Compare commits

..

1 Commits

Author SHA1 Message Date
Rahul Padigela b04f88b222 improvement: new project structure 2019-05-22 22:26:17 -07:00
677 changed files with 21230 additions and 70559 deletions

View File

@ -1,9 +0,0 @@
node_modules
localData/*
localMetadata/*
# Keep the .git/HEAD file in order to properly report version
.git/objects
.github
.tox
coverage
.DS_Store

View File

@ -1,54 +1 @@
{
"extends": "scality",
"plugins": [
"mocha"
],
"rules": {
"import/extensions": "off",
"lines-around-directive": "off",
"no-underscore-dangle": "off",
"indent": "off",
"object-curly-newline": "off",
"operator-linebreak": "off",
"function-paren-newline": "off",
"import/newline-after-import": "off",
"prefer-destructuring": "off",
"implicit-arrow-linebreak": "off",
"no-bitwise": "off",
"dot-location": "off",
"comma-dangle": "off",
"no-undef-init": "off",
"global-require": "off",
"import/no-dynamic-require": "off",
"class-methods-use-this": "off",
"no-plusplus": "off",
"no-else-return": "off",
"object-property-newline": "off",
"import/order": "off",
"no-continue": "off",
"no-tabs": "off",
"lines-between-class-members": "off",
"prefer-spread": "off",
"no-lonely-if": "off",
"no-useless-escape": "off",
"no-restricted-globals": "off",
"no-buffer-constructor": "off",
"import/no-extraneous-dependencies": "off",
"space-unary-ops": "off",
"no-useless-return": "off",
"no-unexpected-multiline": "off",
"no-mixed-operators": "off",
"newline-per-chained-call": "off",
"operator-assignment": "off",
"spaced-comment": "off",
"comma-style": "off",
"no-restricted-properties": "off",
"new-parens": "off",
"no-multi-spaces": "off",
"quote-props": "off",
"mocha/no-exclusive-tests": "error",
},
"parserOptions": {
"ecmaVersion": 2020
}
}
{ "extends": "scality" }

View File

@ -1,32 +1,19 @@
# General support information
# Issue template
GitHub Issues are **reserved** for actionable bug reports (including
documentation inaccuracies), and feature requests.
**All questions** (regarding configuration, use cases, performance, community,
events, setup and usage recommendations, among other things) should be asked on
the **[Zenko Forum](http://forum.zenko.io/)**.
If you are reporting a new issue, make sure that we do not have any
duplicates already open. You can ensure this by searching the issue list for
this repository. If there is a duplicate, please close your issue and add a
comment to the existing issue instead.
> Questions opened as GitHub issues will systematically be closed, and moved to
> the [Zenko Forum](http://forum.zenko.io/).
## General support information
--------------------------------------------------------------------------------
## Avoiding duplicates
When reporting a new issue/requesting a feature, make sure that we do not have
any duplicates already open:
- search the issue list for this repository (use the search bar, select
"Issues" on the left pane after searching);
- if there is a duplicate, please do not open your issue, and add a comment
to the existing issue instead.
--------------------------------------------------------------------------------
GitHub Issues are reserved for actionable bug reports and feature requests.
General questions should be sent to the
[S3 scality server Forum](http://forum.scality.com/).
## Bug report information
(delete this section (everything between the lines) if you're not reporting a bug
but requesting a feature)
(delete this section if not applicable)
### Description
@ -42,22 +29,13 @@ Describe the results you received
### Expected result
Describe the results you expected
Describe the results you expecteds
### Additional information
- Node.js version,
- Docker version,
- yarn version,
- distribution/OS,
- optional: anything else you deem helpful to us.
--------------------------------------------------------------------------------
### Additional information: (Node.js version, Docker version, etc)
## Feature Request
(delete this section (everything between the lines) if you're not requesting
a feature but reporting a bug)
(delete this section if not applicable)
### Proposal
@ -74,14 +52,3 @@ What you would like to happen
### Use case
Please provide use cases for changing the current behavior
### Additional information
- Is this request for your company? Y/N
- If Y: Company name:
- Are you using any Scality Enterprise Edition products (RING, Zenko EE)? Y/N
- Are you willing to contribute this feature yourself?
- Position/Title:
- How did you hear about us?
--------------------------------------------------------------------------------

View File

@ -1,43 +0,0 @@
---
name: "Setup CI environment"
description: "Setup Cloudserver CI environment"
runs:
using: composite
steps:
- name: Setup etc/hosts
shell: bash
run: sudo echo "127.0.0.1 bucketwebsitetester.s3-website-us-east-1.amazonaws.com" | sudo tee -a /etc/hosts
- name: Setup Credentials
shell: bash
run: bash .github/scripts/credentials.bash
- name: Setup job artifacts directory
shell: bash
run: |-
set -exu;
mkdir -p /tmp/artifacts/${JOB_NAME}/;
- uses: actions/setup-node@v4
with:
node-version: '16'
cache: 'yarn'
- name: install dependencies
shell: bash
run: yarn install --ignore-engines --frozen-lockfile --network-concurrency 1
- uses: actions/cache@v3
with:
path: ~/.cache/pip
key: ${{ runner.os }}-pip
- uses: actions/setup-python@v4
with:
python-version: 3.9
- name: Setup python2 test environment
shell: bash
run: |
sudo apt-get install -y libdigest-hmac-perl
pip install 's3cmd==2.3.0'
- name: fix sproxyd.conf permissions
shell: bash
run: sudo chown root:root .github/docker/sproxyd/conf/sproxyd0.conf
- name: ensure fuse kernel module is loaded (for sproxyd)
shell: bash
run: sudo modprobe fuse

View File

@ -1,25 +0,0 @@
FROM ceph/daemon:v3.2.1-stable-3.2-mimic-centos-7
ENV CEPH_DAEMON demo
ENV CEPH_DEMO_DAEMONS mon,mgr,osd,rgw
ENV CEPH_DEMO_UID zenko
ENV CEPH_DEMO_ACCESS_KEY accessKey1
ENV CEPH_DEMO_SECRET_KEY verySecretKey1
ENV CEPH_DEMO_BUCKET zenkobucket
ENV CEPH_PUBLIC_NETWORK 0.0.0.0/0
ENV MON_IP 0.0.0.0
ENV NETWORK_AUTO_DETECT 4
ENV RGW_CIVETWEB_PORT 8001
RUN rm /etc/yum.repos.d/tcmu-runner.repo
ADD ./entrypoint-wrapper.sh /
RUN chmod +x /entrypoint-wrapper.sh && \
yum install -y python-pip && \
yum clean all && \
pip install awscli && \
rm -rf /root/.cache/pip
ENTRYPOINT [ "/entrypoint-wrapper.sh" ]

View File

@ -1,37 +0,0 @@
#!/bin/sh
touch /artifacts/ceph.log
mkfifo /tmp/entrypoint_output
# We run this in the background so that we can tail the RGW log after init,
# because entrypoint.sh never returns
# The next line will be needed when ceph builds 3.2.2 so I'll leave it here
# bash /opt/ceph-container/bin/entrypoint.sh > /tmp/entrypoint_output &
bash /entrypoint.sh > /tmp/entrypoint_output &
entrypoint_pid="$!"
while read -r line; do
echo $line
# When we find this line server has started
if [ -n "$(echo $line | grep 'Creating bucket')" ]; then
break
fi
done < /tmp/entrypoint_output
# Make our buckets - CEPH_DEMO_BUCKET is set to force the "Creating bucket" message, but unused
s3cmd mb s3://cephbucket s3://cephbucket2
mkdir /root/.aws
cat > /root/.aws/credentials <<EOF
[default]
aws_access_key_id = accessKey1
aws_secret_access_key = verySecretKey1
EOF
# Enable versioning on them
for bucket in cephbucket cephbucket2; do
echo "Enabling versiong for $bucket"
aws --endpoint http://127.0.0.1:8001 s3api put-bucket-versioning --bucket $bucket --versioning Status=Enabled
done
tail -f /var/log/ceph/client.rgw.*.log | tee -a /artifacts/ceph.log
wait $entrypoint_pid

View File

@ -1,11 +0,0 @@
#!/bin/sh
# This script is needed because RADOS Gateway
# will open the port before beginning to serve traffic
# causing wait_for_local_port.bash to exit immediately
echo 'Waiting for ceph'
while [ -z "$(curl 127.0.0.1:8001 2>/dev/null)" ]; do
sleep 1
echo -n "."
done

View File

@ -1,10 +0,0 @@
---
version: 2
updates:
- package-ecosystem: npm
directory: "/"
schedule:
interval: daily
time: "13:00"
open-pull-requests-limit: 10
target-branch: "development/7.4"

View File

@ -1,36 +0,0 @@
azurebackend_AZURE_STORAGE_ACCESS_KEY
azurebackend_AZURE_STORAGE_ACCOUNT_NAME
azurebackend_AZURE_STORAGE_ENDPOINT
azurebackend2_AZURE_STORAGE_ACCESS_KEY
azurebackend2_AZURE_STORAGE_ACCOUNT_NAME
azurebackend2_AZURE_STORAGE_ENDPOINT
azurebackendmismatch_AZURE_STORAGE_ACCESS_KEY
azurebackendmismatch_AZURE_STORAGE_ACCOUNT_NAME
azurebackendmismatch_AZURE_STORAGE_ENDPOINT
azurenonexistcontainer_AZURE_STORAGE_ACCESS_KEY
azurenonexistcontainer_AZURE_STORAGE_ACCOUNT_NAME
azurenonexistcontainer_AZURE_STORAGE_ENDPOINT
azuretest_AZURE_BLOB_ENDPOINT
b2backend_B2_ACCOUNT_ID
b2backend_B2_STORAGE_ACCESS_KEY
GOOGLE_SERVICE_EMAIL
GOOGLE_SERVICE_KEY
AWS_S3_BACKEND_ACCESS_KEY
AWS_S3_BACKEND_SECRET_KEY
AWS_S3_BACKEND_ACCESS_KEY_2
AWS_S3_BACKEND_SECRET_KEY_2
AWS_GCP_BACKEND_ACCESS_KEY
AWS_GCP_BACKEND_SECRET_KEY
AWS_GCP_BACKEND_ACCESS_KEY_2
AWS_GCP_BACKEND_SECRET_KEY_2
b2backend_B2_STORAGE_ENDPOINT
gcpbackend2_GCP_SERVICE_EMAIL
gcpbackend2_GCP_SERVICE_KEY
gcpbackend2_GCP_SERVICE_KEYFILE
gcpbackend_GCP_SERVICE_EMAIL
gcpbackend_GCP_SERVICE_KEY
gcpbackendmismatch_GCP_SERVICE_EMAIL
gcpbackendmismatch_GCP_SERVICE_KEY
gcpbackend_GCP_SERVICE_KEYFILE
gcpbackendmismatch_GCP_SERVICE_KEYFILE
gcpbackendnoproxy_GCP_SERVICE_KEYFILE

View File

@ -1,92 +0,0 @@
services:
cloudserver:
image: ${CLOUDSERVER_IMAGE}
command: sh -c "yarn start > /artifacts/s3.log"
network_mode: "host"
volumes:
- /tmp/ssl:/ssl
- /tmp/ssl-kmip:/ssl-kmip
- ${HOME}/.aws/credentials:/root/.aws/credentials
- /tmp/artifacts/${JOB_NAME}:/artifacts
environment:
- CI=true
- ENABLE_LOCAL_CACHE=true
- REDIS_HOST=0.0.0.0
- REDIS_PORT=6379
- REPORT_TOKEN=report-token-1
- REMOTE_MANAGEMENT_DISABLE=1
- HEALTHCHECKS_ALLOWFROM=0.0.0.0/0
- DATA_HOST=0.0.0.0
- METADATA_HOST=0.0.0.0
- S3BACKEND
- S3DATA
- S3METADATA
- MPU_TESTING
- S3VAULT
- S3_LOCATION_FILE
- ENABLE_UTAPI_V2
- BUCKET_DENY_FILTER
- S3KMS
- S3KMIP_PORT
- S3KMIP_HOSTS
- S3KMIP-COMPOUND_CREATE
- S3KMIP_BUCKET_ATTRIBUTE_NAME
- S3KMIP_PIPELINE_DEPTH
- S3KMIP_KEY
- S3KMIP_CERT
- S3KMIP_CA
- MONGODB_HOSTS=0.0.0.0:27018
- MONGODB_RS=rs0
- DEFAULT_BUCKET_KEY_FORMAT
- METADATA_MAX_CACHED_BUCKETS
- ENABLE_NULL_VERSION_COMPAT_MODE
- SCUBA_HOST
- SCUBA_PORT
- SCUBA_HEALTHCHECK_FREQUENCY
- S3QUOTA
- QUOTA_ENABLE_INFLIGHTS
env_file:
- creds.env
depends_on:
- redis
extra_hosts:
- "bucketwebsitetester.s3-website-us-east-1.amazonaws.com:127.0.0.1"
- "pykmip.local:127.0.0.1"
redis:
image: redis:alpine
network_mode: "host"
squid:
network_mode: "host"
profiles: ['ci-proxy']
image: scality/ci-squid
command: >-
sh -c 'mkdir -p /ssl &&
openssl req -new -newkey rsa:2048 -sha256 -days 365 -nodes -x509 \
-subj "/C=US/ST=Country/L=City/O=Organization/CN=CN=scality-proxy" \
-keyout /ssl/myca.pem -out /ssl/myca.pem &&
cp /ssl/myca.pem /ssl/CA.pem &&
squid -f /etc/squid/squid.conf -N -z &&
squid -f /etc/squid/squid.conf -NYCd 1'
volumes:
- /tmp/ssl:/ssl
pykmip:
network_mode: "host"
profiles: ['pykmip']
image: ${PYKMIP_IMAGE:-ghcr.io/scality/cloudserver/pykmip}
volumes:
- /tmp/artifacts/${JOB_NAME}:/artifacts
mongo:
network_mode: "host"
profiles: ['mongo', 'ceph']
image: ${MONGODB_IMAGE}
ceph:
network_mode: "host"
profiles: ['ceph']
image: ghcr.io/scality/cloudserver/ci-ceph
sproxyd:
network_mode: "host"
profiles: ['sproxyd']
image: sproxyd-standalone
build: ./sproxyd
user: 0:0
privileged: yes

View File

@ -1,28 +0,0 @@
FROM mongo:5.0.21
ENV USER=scality \
HOME_DIR=/home/scality \
CONF_DIR=/conf \
DATA_DIR=/data
# Set up directories and permissions
RUN mkdir -p /data/db /data/configdb && chown -R mongodb:mongodb /data/db /data/configdb; \
mkdir /logs; \
adduser --uid 1000 --disabled-password --gecos --quiet --shell /bin/bash scality
# Set up environment variables and directories for scality user
RUN mkdir ${CONF_DIR} && \
chown -R ${USER} ${CONF_DIR} && \
chown -R ${USER} ${DATA_DIR}
# copy the mongo config file
COPY /conf/mongod.conf /conf/mongod.conf
COPY /conf/mongo-run.sh /conf/mongo-run.sh
COPY /conf/initReplicaSet /conf/initReplicaSet.js
EXPOSE 27017/tcp
EXPOSE 27018
# Set up CMD
ENTRYPOINT ["bash", "/conf/mongo-run.sh"]
CMD ["bash", "/conf/mongo-run.sh"]

View File

@ -1,4 +0,0 @@
rs.initiate({
_id: "rs0",
members: [{ _id: 0, host: "127.0.0.1:27018" }]
});

View File

@ -1,10 +0,0 @@
#!/bin/bash
set -exo pipefail
init_RS() {
sleep 5
mongo --port 27018 /conf/initReplicaSet.js
}
init_RS &
mongod --bind_ip_all --config=/conf/mongod.conf

View File

@ -1,15 +0,0 @@
storage:
journal:
enabled: true
engine: wiredTiger
dbPath: "/data/db"
processManagement:
fork: false
net:
port: 27018
bindIp: 0.0.0.0
replication:
replSetName: "rs0"
enableMajorityReadConcern: true
security:
authorization: disabled

View File

@ -1,3 +0,0 @@
FROM ghcr.io/scality/federation/sproxyd:7.10.6.8
ADD ./conf/supervisord.conf ./conf/nginx.conf ./conf/fastcgi_params ./conf/sproxyd0.conf /conf/
RUN chown root:root /conf/sproxyd0.conf

View File

@ -1,26 +0,0 @@
fastcgi_param QUERY_STRING $query_string;
fastcgi_param REQUEST_METHOD $request_method;
fastcgi_param CONTENT_TYPE $content_type;
fastcgi_param CONTENT_LENGTH $content_length;
#fastcgi_param SCRIPT_NAME $fastcgi_script_name;
fastcgi_param SCRIPT_NAME /var/www;
fastcgi_param PATH_INFO $document_uri;
fastcgi_param REQUEST_URI $request_uri;
fastcgi_param DOCUMENT_URI $document_uri;
fastcgi_param DOCUMENT_ROOT $document_root;
fastcgi_param SERVER_PROTOCOL $server_protocol;
fastcgi_param HTTPS $https if_not_empty;
fastcgi_param GATEWAY_INTERFACE CGI/1.1;
fastcgi_param SERVER_SOFTWARE nginx/$nginx_version;
fastcgi_param REMOTE_ADDR $remote_addr;
fastcgi_param REMOTE_PORT $remote_port;
fastcgi_param SERVER_ADDR $server_addr;
fastcgi_param SERVER_PORT $server_port;
fastcgi_param SERVER_NAME $server_name;
# PHP only, required if PHP was built with --enable-force-cgi-redirect
fastcgi_param REDIRECT_STATUS 200;

View File

@ -1,88 +0,0 @@
worker_processes 1;
error_log /logs/error.log;
user root root;
events {
worker_connections 1000;
reuse_port on;
multi_accept on;
}
worker_rlimit_nofile 20000;
http {
root /var/www/;
upstream sproxyds {
least_conn;
keepalive 40;
server 127.0.0.1:20000;
}
server {
client_max_body_size 0;
client_body_timeout 150;
client_header_timeout 150;
postpone_output 0;
client_body_postpone_size 0;
keepalive_requests 1100;
keepalive_timeout 300s;
server_tokens off;
default_type application/octet-stream;
gzip off;
tcp_nodelay on;
tcp_nopush on;
sendfile on;
listen 81;
server_name localhost;
rewrite ^/arc/(.*)$ /dc1/$1 permanent;
location ~* ^/proxy/(.*)$ {
rewrite ^/proxy/(.*)$ /$1 last;
}
allow 127.0.0.1;
deny all;
set $usermd '-';
set $sentusermd '-';
set $elapsed_ms '-';
set $now '-';
log_by_lua '
if not(ngx.var.http_x_scal_usermd == nil) and string.len(ngx.var.http_x_scal_usermd) > 2 then
ngx.var.usermd = string.sub(ngx.decode_base64(ngx.var.http_x_scal_usermd),1,-3)
end
if not(ngx.var.sent_http_x_scal_usermd == nil) and string.len(ngx.var.sent_http_x_scal_usermd) > 2 then
ngx.var.sentusermd = string.sub(ngx.decode_base64(ngx.var.sent_http_x_scal_usermd),1,-3)
end
local elapsed_ms = tonumber(ngx.var.request_time)
if not ( elapsed_ms == nil) then
elapsed_ms = elapsed_ms * 1000
ngx.var.elapsed_ms = tostring(elapsed_ms)
end
local time = tonumber(ngx.var.msec) * 1000
ngx.var.now = time
';
log_format irm '{ "time":"$now","connection":"$connection","request":"$connection_requests","hrtime":"$msec",'
'"httpMethod":"$request_method","httpURL":"$uri","elapsed_ms":$elapsed_ms,'
'"httpCode":$status,"requestLength":$request_length,"bytesSent":$bytes_sent,'
'"contentLength":"$content_length","sentContentLength":"$sent_http_content_length",'
'"contentType":"$content_type","s3Address":"$remote_addr",'
'"requestUserMd":"$usermd","responseUserMd":"$sentusermd",'
'"ringKeyVersion":"$sent_http_x_scal_version","ringStatus":"$sent_http_x_scal_ring_status",'
'"s3Port":"$remote_port","sproxydStatus":"$upstream_status","req_id":"$http_x_scal_request_uids",'
'"ifMatch":"$http_if_match","ifNoneMatch":"$http_if_none_match",'
'"range":"$http_range","contentRange":"$sent_http_content_range","nginxPID":$PID,'
'"sproxydAddress":"$upstream_addr","sproxydResponseTime_s":"$upstream_response_time" }';
access_log /dev/stdout irm;
error_log /dev/stdout error;
location / {
proxy_request_buffering off;
fastcgi_request_buffering off;
fastcgi_no_cache 1;
fastcgi_cache_bypass 1;
fastcgi_buffering off;
fastcgi_ignore_client_abort on;
fastcgi_keep_conn on;
include fastcgi_params;
fastcgi_pass sproxyds;
fastcgi_next_upstream error timeout;
fastcgi_send_timeout 285s;
fastcgi_read_timeout 285s;
}
}
}

View File

@ -1,12 +0,0 @@
{
"general": {
"ring": "DATA",
"port": 20000,
"syslog_facility": "local0"
},
"ring_driver:0": {
"alias": "dc1",
"type": "local",
"queue_path": "/tmp/ring-objs"
},
}

View File

@ -1,43 +0,0 @@
[supervisord]
nodaemon = true
loglevel = info
logfile = %(ENV_LOG_DIR)s/supervisord.log
pidfile = %(ENV_SUP_RUN_DIR)s/supervisord.pid
logfile_maxbytes = 20MB
logfile_backups = 2
[unix_http_server]
file = %(ENV_SUP_RUN_DIR)s/supervisor.sock
[rpcinterface:supervisor]
supervisor.rpcinterface_factory = supervisor.rpcinterface:make_main_rpcinterface
[supervisorctl]
serverurl = unix://%(ENV_SUP_RUN_DIR)s/supervisor.sock
[program:nginx]
directory=%(ENV_SUP_RUN_DIR)s
command=bash -c "/usr/sbin/nginx -c %(ENV_CONF_DIR)s/nginx.conf -g 'daemon off;'"
stdout_logfile = %(ENV_LOG_DIR)s/%(program_name)s-%(process_num)s.log
stderr_logfile = %(ENV_LOG_DIR)s/%(program_name)s-%(process_num)s-stderr.log
stdout_logfile_maxbytes=100MB
stdout_logfile_backups=7
stderr_logfile_maxbytes=100MB
stderr_logfile_backups=7
autorestart=true
autostart=true
user=root
[program:sproxyd]
directory=%(ENV_SUP_RUN_DIR)s
process_name=%(program_name)s-%(process_num)s
numprocs=1
numprocs_start=0
command=/usr/bin/sproxyd -dlw -V127 -c %(ENV_CONF_DIR)s/sproxyd%(process_num)s.conf -P /run%(process_num)s
stdout_logfile = %(ENV_LOG_DIR)s/%(program_name)s-%(process_num)s.log
stdout_logfile_maxbytes=100MB
stdout_logfile_backups=7
redirect_stderr=true
autorestart=true
autostart=true
user=root

View File

@ -1,18 +0,0 @@
-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----

View File

@ -1,18 +0,0 @@
-----BEGIN CERTIFICATE-----
MIIC2zCCAcOgAwIBAgIUIlE8UAkqQ+6mbJDtrt9kkmi8aJYwDQYJKoZIhvcNAQEL
BQAwJDEQMA4GA1UECgwHU2NhbGl0eTEQMA4GA1UEAwwHUm9vdCBDQTAgFw0yMTA0
MDkwMDI4MTFaGA8yMTIxMDMxNjAwMjgxMVowKTEQMA4GA1UECgwHU2NhbGl0eTEV
MBMGA1UEAwwMcHlrbWlwLmxvY2FsMIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIB
CgKCAQEAtxr7pq/lnzVeZz4z52Yc3DeaPqjNfRSyW5cPUlT7ABXFb7+tja7K2C7u
DYVK+Q+2yJCQwYJY47aKJB++ewam9t2V8Xy0Z8S+0I2ImCwuyeihaD/f6uJZRzms
ycdECH22BA6tCPlQLnlboRiZzI6rcIvXAbUMvLvFm3nyYIs9qidExRnfyMjISknM
V+83LT5QW4IcHgKYqzdz2ZmOnk+f4wmMmitcivTdIZCL8Z0cxr7BJlOh5JZ/V5uj
WUXeNa+ttW0RKKBlg9T+wj0JvwoJBPZTmsMAy3tI9tjLg3DwGYKsflbFeU2tebXI
gncGFZ/dFxj331GGtq3kz1PzAUYf2wIDAQABMA0GCSqGSIb3DQEBCwUAA4IBAQB1
8HgJ0fu6/pCrDxAm90eESFjmaTFyTN8q00zhq4Cb3zAT9KMWzAygkZ9n4ZFgELPo
7kBE2H6RcDdoBmjVYd8HnBloDdYzYbncKgt5YBvxRaMSF4/l65BM8wjatyXErqnH
QLLTRe5AuF0/F0KtPeDQ2JFVu8dZ35W3fyKGPRsEdVOSCTHROmqpGhZCpscyUP4W
Hb0dBTESQ9mQHw14OCaaahARd0X5WdcA/E+m0fpGqj1rQCXS+PrRcSLe1E1hqPlK
q/hXSXD5nybwipktELvJCbB7l4HmJr2pIpldeR5+ef68Cs8hqs6DRlsJX9sK2ng+
TFe5v6SCarqZ9kFvr6Yp
-----END CERTIFICATE-----

View File

@ -1,18 +0,0 @@
-----BEGIN CERTIFICATE-----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-----END CERTIFICATE-----

View File

@ -1,28 +0,0 @@
-----BEGIN PRIVATE KEY-----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-----END PRIVATE KEY-----

View File

@ -1,28 +0,0 @@
-----BEGIN PRIVATE KEY-----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-----END PRIVATE KEY-----

View File

@ -1,35 +0,0 @@
name: Test alerts
on:
push:
branches-ignore:
- 'development/**'
- 'q/*/**'
jobs:
run-alert-tests:
runs-on: ubuntu-latest
strategy:
matrix:
tests:
- name: 1 minute interval tests
file: monitoring/alerts.test.yaml
- name: 10 seconds interval tests
file: monitoring/alerts.10s.test.yaml
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Render and test ${{ matrix.tests.name }}
uses: scality/action-prom-render-test@1.0.3
with:
alert_file_path: monitoring/alerts.yaml
test_file_path: ${{ matrix.tests.file }}
alert_inputs: |
namespace=zenko
service=artesca-data-connector-s3api-metrics
reportJob=artesca-data-ops-report-handler
replicas=3
github_token: ${{ secrets.GITHUB_TOKEN }}

View File

@ -1,25 +0,0 @@
---
name: codeQL
on:
push:
branches: [w/**, q/*]
pull_request:
branches: [development/*, stabilization/*, hotfix/*]
workflow_dispatch:
jobs:
analyze:
name: Static analysis with CodeQL
runs-on: ubuntu-latest
steps:
- name: Checkout code
uses: actions/checkout@v4
- name: Initialize CodeQL
uses: github/codeql-action/init@v3
with:
languages: javascript, python, ruby
- name: Build and analyze
uses: github/codeql-action/analyze@v3

View File

@ -1,16 +0,0 @@
---
name: dependency review
on:
pull_request:
branches: [development/*, stabilization/*, hotfix/*]
jobs:
dependency-review:
runs-on: ubuntu-latest
steps:
- name: 'Checkout Repository'
uses: actions/checkout@v4
- name: 'Dependency Review'
uses: actions/dependency-review-action@v4

View File

@ -1,80 +0,0 @@
---
name: release
run-name: release ${{ inputs.tag }}
on:
workflow_dispatch:
inputs:
tag:
description: 'Tag to be released'
required: true
env:
PROJECT_NAME: ${{ github.event.repository.name }}
jobs:
build-federation-image:
runs-on: ubuntu-20.04
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Login to GitHub Registry
uses: docker/login-action@v3
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ github.token }}
- name: Build and push image for federation
uses: docker/build-push-action@v5
with:
push: true
context: .
file: images/svc-base/Dockerfile
tags: |
ghcr.io/${{ github.repository }}:${{ github.event.inputs.tag }}-svc-base
cache-from: type=gha,scope=federation
cache-to: type=gha,mode=max,scope=federation
release:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Set up Docker Buildk
uses: docker/setup-buildx-action@v3
- name: Login to Registry
uses: docker/login-action@v3
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ github.token }}
- name: Push dashboards into the production namespace
run: |
oras push ghcr.io/${{ github.repository }}/${{ env.PROJECT_NAME }}-dashboards:${{ github.event.inputs.tag }} \
dashboard.json:application/grafana-dashboard+json \
alerts.yaml:application/prometheus-alerts+yaml
working-directory: monitoring
- name: Build and push
uses: docker/build-push-action@v5
with:
context: .
push: true
tags: ghcr.io/${{ github.repository }}:${{ github.event.inputs.tag }}
cache-from: type=gha
cache-to: type=gha,mode=max
- name: Create Release
uses: softprops/action-gh-release@v2
env:
GITHUB_TOKEN: ${{ github.token }}
with:
name: Release ${{ github.event.inputs.tag }}
tag_name: ${{ github.event.inputs.tag }}
generate_release_notes: true
target_commitish: ${{ github.sha }}

View File

@ -1,533 +0,0 @@
---
name: tests
on:
workflow_dispatch:
push:
branches-ignore:
- 'development/**'
- 'q/*/**'
env:
# Secrets
azurebackend_AZURE_STORAGE_ACCESS_KEY: >-
${{ secrets.AZURE_STORAGE_ACCESS_KEY }}
azurebackend_AZURE_STORAGE_ACCOUNT_NAME: >-
${{ secrets.AZURE_STORAGE_ACCOUNT_NAME }}
azurebackend_AZURE_STORAGE_ENDPOINT: >-
${{ secrets.AZURE_STORAGE_ENDPOINT }}
azurebackend2_AZURE_STORAGE_ACCESS_KEY: >-
${{ secrets.AZURE_STORAGE_ACCESS_KEY_2 }}
azurebackend2_AZURE_STORAGE_ACCOUNT_NAME: >-
${{ secrets.AZURE_STORAGE_ACCOUNT_NAME_2 }}
azurebackend2_AZURE_STORAGE_ENDPOINT: >-
${{ secrets.AZURE_STORAGE_ENDPOINT_2 }}
azurebackendmismatch_AZURE_STORAGE_ACCESS_KEY: >-
${{ secrets.AZURE_STORAGE_ACCESS_KEY }}
azurebackendmismatch_AZURE_STORAGE_ACCOUNT_NAME: >-
${{ secrets.AZURE_STORAGE_ACCOUNT_NAME }}
azurebackendmismatch_AZURE_STORAGE_ENDPOINT: >-
${{ secrets.AZURE_STORAGE_ENDPOINT }}
azurenonexistcontainer_AZURE_STORAGE_ACCESS_KEY: >-
${{ secrets.AZURE_STORAGE_ACCESS_KEY }}
azurenonexistcontainer_AZURE_STORAGE_ACCOUNT_NAME: >-
${{ secrets.AZURE_STORAGE_ACCOUNT_NAME }}
azurenonexistcontainer_AZURE_STORAGE_ENDPOINT: >-
${{ secrets.AZURE_STORAGE_ENDPOINT }}
azuretest_AZURE_BLOB_ENDPOINT: "${{ secrets.AZURE_STORAGE_ENDPOINT }}"
b2backend_B2_ACCOUNT_ID: "${{ secrets.B2BACKEND_B2_ACCOUNT_ID }}"
b2backend_B2_STORAGE_ACCESS_KEY: >-
${{ secrets.B2BACKEND_B2_STORAGE_ACCESS_KEY }}
GOOGLE_SERVICE_EMAIL: "${{ secrets.GCP_SERVICE_EMAIL }}"
GOOGLE_SERVICE_KEY: "${{ secrets.GCP_SERVICE_KEY }}"
AWS_S3_BACKEND_ACCESS_KEY: "${{ secrets.AWS_S3_BACKEND_ACCESS_KEY }}"
AWS_S3_BACKEND_SECRET_KEY: "${{ secrets.AWS_S3_BACKEND_SECRET_KEY }}"
AWS_S3_BACKEND_ACCESS_KEY_2: "${{ secrets.AWS_S3_BACKEND_ACCESS_KEY_2 }}"
AWS_S3_BACKEND_SECRET_KEY_2: "${{ secrets.AWS_S3_BACKEND_SECRET_KEY_2 }}"
AWS_GCP_BACKEND_ACCESS_KEY: "${{ secrets.AWS_GCP_BACKEND_ACCESS_KEY }}"
AWS_GCP_BACKEND_SECRET_KEY: "${{ secrets.AWS_GCP_BACKEND_SECRET_KEY }}"
AWS_GCP_BACKEND_ACCESS_KEY_2: "${{ secrets.AWS_GCP_BACKEND_ACCESS_KEY_2 }}"
AWS_GCP_BACKEND_SECRET_KEY_2: "${{ secrets.AWS_GCP_BACKEND_SECRET_KEY_2 }}"
b2backend_B2_STORAGE_ENDPOINT: "${{ secrets.B2BACKEND_B2_STORAGE_ENDPOINT }}"
gcpbackend2_GCP_SERVICE_EMAIL: "${{ secrets.GCP2_SERVICE_EMAIL }}"
gcpbackend2_GCP_SERVICE_KEY: "${{ secrets.GCP2_SERVICE_KEY }}"
gcpbackend2_GCP_SERVICE_KEYFILE: /root/.gcp/servicekey
gcpbackend_GCP_SERVICE_EMAIL: "${{ secrets.GCP_SERVICE_EMAIL }}"
gcpbackend_GCP_SERVICE_KEY: "${{ secrets.GCP_SERVICE_KEY }}"
gcpbackendmismatch_GCP_SERVICE_EMAIL: >-
${{ secrets.GCPBACKENDMISMATCH_GCP_SERVICE_EMAIL }}
gcpbackendmismatch_GCP_SERVICE_KEY: >-
${{ secrets.GCPBACKENDMISMATCH_GCP_SERVICE_KEY }}
gcpbackend_GCP_SERVICE_KEYFILE: /root/.gcp/servicekey
gcpbackendmismatch_GCP_SERVICE_KEYFILE: /root/.gcp/servicekey
gcpbackendnoproxy_GCP_SERVICE_KEYFILE: /root/.gcp/servicekey
gcpbackendproxy_GCP_SERVICE_KEYFILE: /root/.gcp/servicekey
# Configs
ENABLE_LOCAL_CACHE: "true"
REPORT_TOKEN: "report-token-1"
REMOTE_MANAGEMENT_DISABLE: "1"
# https://github.com/git-lfs/git-lfs/issues/5749
GIT_CLONE_PROTECTION_ACTIVE: 'false'
jobs:
linting-coverage:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v4
- uses: actions/setup-node@v4
with:
node-version: '16'
cache: yarn
- name: install dependencies
run: yarn install --frozen-lockfile --network-concurrency 1
- uses: actions/setup-python@v5
with:
python-version: '3.9'
- uses: actions/cache@v4
with:
path: ~/.cache/pip
key: ${{ runner.os }}-pip
- name: Install python deps
run: pip install flake8
- name: Lint Javascript
run: yarn run --silent lint -- --max-warnings 0
- name: Lint Markdown
run: yarn run --silent lint_md
- name: Lint python
run: flake8 $(git ls-files "*.py")
- name: Lint Yaml
run: yamllint -c yamllint.yml $(git ls-files "*.yml")
- name: Unit Coverage
run: |
set -ex
mkdir -p $CIRCLE_TEST_REPORTS/unit
yarn test
yarn run test_legacy_location
env:
S3_LOCATION_FILE: tests/locationConfig/locationConfigTests.json
CIRCLE_TEST_REPORTS: /tmp
CIRCLE_ARTIFACTS: /tmp
CI_REPORTS: /tmp
- name: Unit Coverage logs
run: find /tmp/unit -exec cat {} \;
- name: preparing junit files for upload
run: |
mkdir -p artifacts/junit
find . -name "*junit*.xml" -exec cp {} artifacts/junit/ ";"
if: always()
- name: Upload files to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: artifacts
if: always()
build:
runs-on: ubuntu-20.04
permissions:
contents: read
packages: write
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Login to GitHub Registry
uses: docker/login-action@v3
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ github.token }}
- name: Build and push cloudserver image
uses: docker/build-push-action@v5
with:
push: true
context: .
provenance: false
tags: |
ghcr.io/${{ github.repository }}:${{ github.sha }}
labels: |
git.repository=${{ github.repository }}
git.commit-sha=${{ github.sha }}
cache-from: type=gha,scope=cloudserver
cache-to: type=gha,mode=max,scope=cloudserver
- name: Build and push pykmip image
uses: docker/build-push-action@v5
with:
push: true
context: .github/pykmip
tags: |
ghcr.io/${{ github.repository }}/pykmip:${{ github.sha }}
labels: |
git.repository=${{ github.repository }}
git.commit-sha=${{ github.sha }}
cache-from: type=gha,scope=pykmip
cache-to: type=gha,mode=max,scope=pykmip
- name: Build and push MongoDB
uses: docker/build-push-action@v5
with:
push: true
context: .github/docker/mongodb
tags: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
cache-from: type=gha,scope=mongodb
cache-to: type=gha,mode=max,scope=mongodb
multiple-backend:
runs-on: ubuntu-latest
needs: build
env:
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
S3BACKEND: mem
S3_LOCATION_FILE: /usr/src/app/tests/locationConfig/locationConfigTests.json
S3DATA: multiple
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Login to Registry
uses: docker/login-action@v3
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ github.token }}
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Setup CI services
run: docker compose --profile sproxyd up -d
working-directory: .github/docker
- name: Run multiple backend test
run: |-
set -o pipefail;
bash wait_for_local_port.bash 8000 40
bash wait_for_local_port.bash 81 40
yarn run multiple_backend_test | tee /tmp/artifacts/${{ github.job }}/tests.log
env:
S3_LOCATION_FILE: tests/locationConfig/locationConfigTests.json
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
mongo-v0-ft-tests:
runs-on: ubuntu-latest
needs: build
env:
S3BACKEND: mem
MPU_TESTING: "yes"
S3METADATA: mongodb
S3KMS: file
S3_LOCATION_FILE: /usr/src/app/tests/locationConfig/locationConfigTests.json
DEFAULT_BUCKET_KEY_FORMAT: v0
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Setup CI services
run: docker compose --profile mongo up -d
working-directory: .github/docker
- name: Run functional tests
run: |-
set -o pipefail;
bash wait_for_local_port.bash 8000 40
yarn run ft_test | tee /tmp/artifacts/${{ github.job }}/tests.log
env:
S3_LOCATION_FILE: tests/locationConfig/locationConfigTests.json
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
mongo-v1-ft-tests:
runs-on: ubuntu-latest
needs: build
env:
S3BACKEND: mem
MPU_TESTING: "yes"
S3METADATA: mongodb
S3KMS: file
S3_LOCATION_FILE: /usr/src/app/tests/locationConfig/locationConfigTests.json
DEFAULT_BUCKET_KEY_FORMAT: v1
METADATA_MAX_CACHED_BUCKETS: 1
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Setup CI services
run: docker compose --profile mongo up -d
working-directory: .github/docker
- name: Run functional tests
run: |-
set -o pipefail;
bash wait_for_local_port.bash 8000 40
yarn run ft_test | tee /tmp/artifacts/${{ github.job }}/tests.log
yarn run ft_mixed_bucket_format_version | tee /tmp/artifacts/${{ github.job }}/mixed-tests.log
env:
S3_LOCATION_FILE: tests/locationConfig/locationConfigTests.json
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
file-ft-tests:
strategy:
matrix:
include:
- job-name: file-ft-tests
name: ${{ matrix.job-name }}
runs-on: ubuntu-latest
needs: build
env:
S3BACKEND: file
S3VAULT: mem
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
MPU_TESTING: "yes"
JOB_NAME: ${{ matrix.job-name }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Setup matrix job artifacts directory
shell: bash
run: |
set -exu
mkdir -p /tmp/artifacts/${{ matrix.job-name }}/
- name: Setup CI services
run: docker compose up -d
working-directory: .github/docker
- name: Run file ft tests
run: |-
set -o pipefail;
bash wait_for_local_port.bash 8000 40
yarn run ft_test | tee /tmp/artifacts/${{ matrix.job-name }}/tests.log
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
utapi-v2-tests:
runs-on: ubuntu-latest
needs: build
env:
ENABLE_UTAPI_V2: t
S3BACKEND: mem
BUCKET_DENY_FILTER: utapi-event-filter-deny-bucket
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Setup CI services
run: docker compose up -d
working-directory: .github/docker
- name: Run file utapi v2 tests
run: |-
set -ex -o pipefail;
bash wait_for_local_port.bash 8000 40
yarn run test_utapi_v2 | tee /tmp/artifacts/${{ github.job }}/tests.log
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
quota-tests:
runs-on: ubuntu-latest
needs: build
strategy:
matrix:
inflights:
- name: "With Inflights"
value: "true"
- name: "Without Inflights"
value: "false"
env:
S3METADATA: mongodb
S3BACKEND: mem
S3QUOTA: scuba
QUOTA_ENABLE_INFLIGHTS: ${{ matrix.inflights.value }}
SCUBA_HOST: localhost
SCUBA_PORT: 8100
SCUBA_HEALTHCHECK_FREQUENCY: 100
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Setup CI services
run: docker compose --profile mongo up -d
working-directory: .github/docker
- name: Run quota tests
run: |-
set -ex -o pipefail;
bash wait_for_local_port.bash 8000 40
yarn run test_quota | tee /tmp/artifacts/${{ github.job }}/tests.log
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
kmip-ft-tests:
runs-on: ubuntu-latest
needs: build
env:
S3BACKEND: file
S3VAULT: mem
MPU_TESTING: "yes"
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
PYKMIP_IMAGE: ghcr.io/${{ github.repository }}/pykmip:${{ github.sha }}
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- name: Copy KMIP certs
run: cp -r ./certs /tmp/ssl-kmip
working-directory: .github/pykmip
- name: Setup CI services
run: docker compose --profile pykmip up -d
working-directory: .github/docker
- name: Run file KMIP tests
run: |-
set -ex -o pipefail;
bash wait_for_local_port.bash 8000 40
bash wait_for_local_port.bash 5696 40
yarn run ft_kmip | tee /tmp/artifacts/${{ github.job }}/tests.log
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()
ceph-backend-test:
runs-on: ubuntu-latest
needs: build
env:
S3BACKEND: mem
S3DATA: multiple
S3KMS: file
CI_CEPH: 'true'
MPU_TESTING: "yes"
S3_LOCATION_FILE: /usr/src/app/tests/locationConfig/locationConfigCeph.json
MONGODB_IMAGE: ghcr.io/${{ github.repository }}/ci-mongodb:${{ github.sha }}
CLOUDSERVER_IMAGE: ghcr.io/${{ github.repository }}:${{ github.sha }}
JOB_NAME: ${{ github.job }}
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Login to GitHub Registry
uses: docker/login-action@v3
with:
registry: ghcr.io
username: ${{ github.repository_owner }}
password: ${{ github.token }}
- name: Setup CI environment
uses: ./.github/actions/setup-ci
- uses: ruby/setup-ruby@v1
with:
ruby-version: '2.5.9'
- name: Install Ruby dependencies
run: |
gem install nokogiri:1.12.5 excon:0.109.0 fog-aws:1.3.0 json mime-types:3.1 rspec:3.5
- name: Install Java dependencies
run: |
sudo apt-get update && sudo apt-get install -y --fix-missing default-jdk maven
- name: Setup CI services
run: docker compose --profile ceph up -d
working-directory: .github/docker
env:
S3METADATA: mongodb
- name: Run Ceph multiple backend tests
run: |-
set -ex -o pipefail;
bash .github/ceph/wait_for_ceph.sh
bash wait_for_local_port.bash 27018 40
bash wait_for_local_port.bash 8000 40
yarn run multiple_backend_test | tee /tmp/artifacts/${{ github.job }}/multibackend-tests.log
env:
S3_LOCATION_FILE: tests/locationConfig/locationConfigTests.json
S3METADATA: mem
- name: Run Java tests
run: |-
set -ex -o pipefail;
mvn test | tee /tmp/artifacts/${{ github.job }}/java-tests.log
working-directory: tests/functional/jaws
- name: Run Ruby tests
run: |-
set -ex -o pipefail;
rspec -fd --backtrace tests.rb | tee /tmp/artifacts/${{ github.job }}/ruby-tests.log
working-directory: tests/functional/fog
- name: Run Javascript AWS SDK tests
run: |-
set -ex -o pipefail;
yarn run ft_awssdk | tee /tmp/artifacts/${{ github.job }}/js-awssdk-tests.log;
yarn run ft_s3cmd | tee /tmp/artifacts/${{ github.job }}/js-s3cmd-tests.log;
env:
S3_LOCATION_FILE: tests/locationConfig/locationConfigCeph.json
S3BACKEND: file
S3VAULT: mem
S3METADATA: mongodb
- name: Upload logs to artifacts
uses: scality/action-artifacts@v4
with:
method: upload
url: https://artifacts.scality.net
user: ${{ secrets.ARTIFACTS_USER }}
password: ${{ secrets.ARTIFACTS_PASSWORD }}
source: /tmp/artifacts
if: always()

8
.gitignore vendored
View File

@ -22,14 +22,6 @@ coverage
# Compiled binary addons (http://nodejs.org/api/addons.html)
build/Release
# Sphinx build dir
_build
# Dependency directory
# https://www.npmjs.org/doc/misc/npm-faq.html#should-i-check-my-node_modules-folder-into-git
node_modules
yarn.lock
.tox
# Junit directory
junit

View File

@ -1,60 +0,0 @@
ARG NODE_VERSION=16.20-bullseye-slim
FROM node:${NODE_VERSION} as builder
WORKDIR /usr/src/app
RUN apt-get update \
&& apt-get install -y --no-install-recommends \
build-essential \
ca-certificates \
curl \
git \
gnupg2 \
jq \
python3 \
ssh \
wget \
libffi-dev \
zlib1g-dev \
&& apt-get clean \
&& mkdir -p /root/ssh \
&& ssh-keyscan -H github.com > /root/ssh/known_hosts
ENV PYTHON=python3
COPY package.json yarn.lock /usr/src/app/
RUN npm install typescript -g
RUN yarn install --production --ignore-optional --frozen-lockfile --ignore-engines --network-concurrency 1
################################################################################
FROM node:${NODE_VERSION}
RUN apt-get update && \
apt-get install -y --no-install-recommends \
jq \
&& rm -rf /var/lib/apt/lists/*
ENV NO_PROXY localhost,127.0.0.1
ENV no_proxy localhost,127.0.0.1
EXPOSE 8000
EXPOSE 8002
RUN apt-get update && \
apt-get install -y --no-install-recommends \
jq \
tini \
&& rm -rf /var/lib/apt/lists/*
WORKDIR /usr/src/app
# Keep the .git directory in order to properly report version
COPY . /usr/src/app
COPY --from=builder /usr/src/app/node_modules ./node_modules/
VOLUME ["/usr/src/app/localData","/usr/src/app/localMetadata"]
ENTRYPOINT ["tini", "--", "/usr/src/app/docker-entrypoint.sh"]
CMD [ "yarn", "start" ]

165
README.md
View File

@ -1,7 +1,12 @@
# Zenko CloudServer with Vitastor Backend
# Zenko CloudServer
![Zenko CloudServer logo](res/scality-cloudserver-logo.png)
[![CircleCI][badgepub]](https://circleci.com/gh/scality/S3)
[![Scality CI][badgepriv]](http://ci.ironmann.io/gh/scality/S3)
[![Docker Pulls][badgedocker]](https://hub.docker.com/r/scality/s3server/)
[![Docker Pulls][badgetwitter]](https://twitter.com/zenko)
## Overview
CloudServer (formerly S3 Server) is an open-source Amazon S3-compatible
@ -11,71 +16,125 @@ Scalitys Open Source Multi-Cloud Data Controller.
CloudServer provides a single AWS S3 API interface to access multiple
backend data storage both on-premise or public in the cloud.
This repository contains a fork of CloudServer with [Vitastor](https://git.yourcmc.ru/vitalif/vitastor)
backend support.
CloudServer is useful for Developers, either to run as part of a
continous integration test environment to emulate the AWS S3 service locally
or as an abstraction layer to develop object storage enabled
application on the go.
## Quick Start with Vitastor
## Learn more at [www.zenko.io/cloudserver](https://www.zenko.io/cloudserver/)
Vitastor Backend is in experimental status, however you can already try to
run it and write or read something, or even mount it with [GeeseFS](https://github.com/yandex-cloud/geesefs),
it works too 😊.
## [May I offer you some lovely documentation?](http://s3-server.readthedocs.io/en/latest/)
Installation instructions:
## Docker
### Install Vitastor
[Run your Zenko CloudServer with Docker](https://hub.docker.com/r/scality/s3server/)
Refer to [Vitastor Quick Start Manual](https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/docs/intro/quickstart.en.md).
## Contributing
### Install Zenko with Vitastor Backend
In order to contribute, please follow the
[Contributing Guidelines](
https://github.com/scality/Guidelines/blob/master/CONTRIBUTING.md).
- Clone this repository: `git clone https://git.yourcmc.ru/vitalif/zenko-cloudserver-vitastor`
- Install dependencies: `npm install --omit dev` or just `npm install`
- Clone Vitastor repository: `git clone https://git.yourcmc.ru/vitalif/vitastor`
- Build Vitastor node.js binding by running `npm install` in `node-binding` subdirectory of Vitastor repository.
You need `node-gyp` and `vitastor-client-dev` (Vitastor client library) for it to succeed.
- Symlink Vitastor module to Zenko: `ln -s /path/to/vitastor/node-binding /path/to/zenko/node_modules/vitastor`
## Installation
### Install and Configure MongoDB
### Dependencies
Refer to [MongoDB Manual](https://www.mongodb.com/docs/manual/installation/).
Building and running the Zenko CloudServer requires node.js 6.9.5 and npm v3
. Up-to-date versions can be found at
[Nodesource](https://github.com/nodesource/distributions).
### Setup Zenko
### Clone source code
- Create a separate pool for S3 object data in your Vitastor cluster: `vitastor-cli create-pool s3-data`
- Retrieve ID of the new pool from `vitastor-cli ls-pools --detail s3-data`
- In another pool, create an image for storing Vitastor volume metadata: `vitastor-cli create -s 10G s3-volume-meta`
- Copy `config.json.vitastor` to `config.json`, adjust it to match your domain
- Copy `authdata.json.example` to `authdata.json` - this is where you set S3 access & secret keys,
and also adjust them if you want to. Scality seems to use a separate auth service "Scality Vault" for
access keys, but it's not published, so let's use a file for now.
- Copy `locationConfig.json.vitastor` to `locationConfig.json` - this is where you set Vitastor cluster access data.
You should put correct values for `pool_id` (pool ID from the second step) and `metadata_image` (from the third step)
in this file.
Note: `locationConfig.json` in this version corresponds to storage classes (like STANDARD, COLD, etc)
instead of "locations" (zones like us-east-1) as it was in original Zenko CloudServer.
### Start Zenko
Start the S3 server with: `node index.js`
If you use default settings, Zenko CloudServer starts on port 8000.
The default access key is `accessKey1` with a secret key of `verySecretKey1`.
Now you can access your S3 with `s3cmd` or `geesefs`:
```
s3cmd --access_key=accessKey1 --secret_key=verySecretKey1 --host=http://localhost:8000 mb s3://testbucket
```shell
git clone https://github.com/scality/S3.git
```
```
AWS_ACCESS_KEY_ID=accessKey1 \
AWS_SECRET_ACCESS_KEY=verySecretKey1 \
geesefs --endpoint http://localhost:8000 testbucket mountdir
### Install js dependencies
Go to the ./S3 folder,
```shell
npm install
```
# Author & License
If you get an error regarding installation of the diskUsage module,
please install g++.
- [Zenko CloudServer](https://s3-server.readthedocs.io/en/latest/) author is Scality, licensed under [Apache License, version 2.0](https://www.apache.org/licenses/LICENSE-2.0)
- [Vitastor](https://git.yourcmc.ru/vitalif/vitastor/) and Zenko Vitastor backend author is Vitaliy Filippov, licensed under [VNPL-1.1](https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/VNPL-1.1.txt)
(a "network copyleft" license based on AGPL/SSPL, but worded in a better way)
If you get an error regarding level-down bindings, try clearing your npm cache:
```shell
npm cache clear
```
## Run it with a file backend
```shell
npm start
```
This starts a Zenko CloudServer on port 8000. Two additional ports 9990 and
9991 are also open locally for internal transfer of metadata and data,
respectively.
The default access key is accessKey1 with
a secret key of verySecretKey1.
By default the metadata files will be saved in the
localMetadata directory and the data files will be saved
in the localData directory within the ./S3 directory on your
machine. These directories have been pre-created within the
repository. If you would like to save the data or metadata in
different locations of your choice, you must specify them with absolute paths.
So, when starting the server:
```shell
mkdir -m 700 $(pwd)/myFavoriteDataPath
mkdir -m 700 $(pwd)/myFavoriteMetadataPath
export S3DATAPATH="$(pwd)/myFavoriteDataPath"
export S3METADATAPATH="$(pwd)/myFavoriteMetadataPath"
npm start
```
## Run it with multiple data backends
```shell
export S3DATA='multiple'
npm start
```
This starts a Zenko CloudServer on port 8000.
The default access key is accessKey1 with
a secret key of verySecretKey1.
With multiple backends, you have the ability to
choose where each object will be saved by setting
the following header with a locationConstraint on
a PUT request:
```shell
'x-amz-meta-scal-location-constraint':'myLocationConstraint'
```
If no header is sent with a PUT object request, the
location constraint of the bucket will determine
where the data is saved. If the bucket has no location
constraint, the endpoint of the PUT request will be
used to determine location.
See the Configuration section below to learn how to set
location constraints.
## Run it with an in-memory backend
```shell
npm run mem_backend
```
This starts a Zenko CloudServer on port 8000.
The default access key is accessKey1 with
a secret key of verySecretKey1.
[badgetwitter]: https://img.shields.io/twitter/follow/zenko.svg?style=social&label=Follow
[badgedocker]: https://img.shields.io/docker/pulls/scality/s3server.svg
[badgepub]: https://circleci.com/gh/scality/S3.svg?style=svg
[badgepriv]: http://ci.ironmann.io/gh/scality/S3.svg?style=svg&circle-token=1f105b7518b53853b5b7cf72302a3f75d8c598ae

View File

@ -1,2 +0,0 @@
---
theme: jekyll-theme-modernist

View File

@ -1,56 +0,0 @@
{
"accounts": [{
"name": "Bart",
"email": "sampleaccount1@sampling.com",
"arn": "arn:aws:iam::123456789012:root",
"canonicalID": "79a59df900b949e55d96a1e698fbacedfd6e09d98eacf8f8d5218e7cd47ef2be",
"shortid": "123456789012",
"keys": [{
"access": "accessKey1",
"secret": "verySecretKey1"
}]
}, {
"name": "Lisa",
"email": "sampleaccount2@sampling.com",
"arn": "arn:aws:iam::123456789013:root",
"canonicalID": "79a59df900b949e55d96a1e698fbacedfd6e09d98eacf8f8d5218e7cd47ef2bf",
"shortid": "123456789013",
"keys": [{
"access": "accessKey2",
"secret": "verySecretKey2"
}]
},
{
"name": "Clueso",
"email": "inspector@clueso.info",
"arn": "arn:aws:iam::123456789014:root",
"canonicalID": "http://acs.zenko.io/accounts/service/clueso",
"shortid": "123456789014",
"keys": [{
"access": "cluesoKey1",
"secret": "cluesoSecretKey1"
}]
},
{
"name": "Replication",
"email": "inspector@replication.info",
"arn": "arn:aws:iam::123456789015:root",
"canonicalID": "http://acs.zenko.io/accounts/service/replication",
"shortid": "123456789015",
"keys": [{
"access": "replicationKey1",
"secret": "replicationSecretKey1"
}]
},
{
"name": "Lifecycle",
"email": "inspector@lifecycle.info",
"arn": "arn:aws:iam::123456789016:root",
"canonicalID": "http://acs.zenko.io/accounts/service/lifecycle",
"shortid": "123456789016",
"keys": [{
"access": "lifecycleKey1",
"secret": "lifecycleSecretKey1"
}]
}]
}

554
backends/data/external/AwsClient.js vendored Normal file
View File

@ -0,0 +1,554 @@
const { errors, s3middleware } = require('arsenal');
const AWS = require('aws-sdk');
const MD5Sum = s3middleware.MD5Sum;
const getMetaHeaders = s3middleware.userMetadata.getMetaHeaders;
const createLogger = require('../multipleBackendLogger');
const { prepareStream } = require('../../api/apiUtils/object/prepareStream');
const { logHelper, removeQuotes, trimXMetaPrefix } = require('./utils');
const { config } = require('../../Config');
const missingVerIdInternalError = errors.InternalError.customizeDescription(
'Invalid state. Please ensure versioning is enabled ' +
'in AWS for the location constraint and try again.'
);
class AwsClient {
constructor(config) {
this.clientType = 'aws_s3';
this._s3Params = config.s3Params;
this._awsBucketName = config.bucketName;
this._bucketMatch = config.bucketMatch;
this._dataStoreName = config.dataStoreName;
this._serverSideEncryption = config.serverSideEncryption;
this._client = new AWS.S3(this._s3Params);
}
_createAwsKey(requestBucketName, requestObjectKey,
bucketMatch) {
if (bucketMatch) {
return requestObjectKey;
}
return `${requestBucketName}/${requestObjectKey}`;
}
put(stream, size, keyContext, reqUids, callback) {
const awsKey = this._createAwsKey(keyContext.bucketName,
keyContext.objectKey, this._bucketMatch);
const metaHeaders = trimXMetaPrefix(keyContext.metaHeaders);
const log = createLogger(reqUids);
const putCb = (err, data) => {
if (err) {
logHelper(log, 'error', 'err from data backend',
err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
if (!data.VersionId) {
logHelper(log, 'error', 'missing version id for data ' +
'backend object', missingVerIdInternalError,
this._dataStoreName);
return callback(missingVerIdInternalError);
}
const dataStoreVersionId = data.VersionId;
return callback(null, awsKey, dataStoreVersionId);
};
const params = {
Bucket: this._awsBucketName,
Key: awsKey,
};
// we call data.put to create a delete marker, but it's actually a
// delete request in call to AWS
if (keyContext.isDeleteMarker) {
return this._client.deleteObject(params, putCb);
}
const uploadParams = params;
uploadParams.Metadata = metaHeaders;
uploadParams.ContentLength = size;
if (this._serverSideEncryption) {
uploadParams.ServerSideEncryption = 'AES256';
}
if (keyContext.tagging) {
uploadParams.Tagging = keyContext.tagging;
}
if (keyContext.contentType !== undefined) {
uploadParams.ContentType = keyContext.contentType;
}
if (keyContext.cacheControl !== undefined) {
uploadParams.CacheControl = keyContext.cacheControl;
}
if (keyContext.contentDisposition !== undefined) {
uploadParams.ContentDisposition = keyContext.contentDisposition;
}
if (keyContext.contentEncoding !== undefined) {
uploadParams.ContentEncoding = keyContext.contentEncoding;
}
if (!stream) {
return this._client.putObject(uploadParams, putCb);
}
uploadParams.Body = stream;
return this._client.upload(uploadParams, putCb);
}
head(objectGetInfo, reqUids, callback) {
const log = createLogger(reqUids);
const { key, dataStoreVersionId } = objectGetInfo;
return this._client.headObject({
Bucket: this._awsBucketName,
Key: key,
VersionId: dataStoreVersionId,
}, err => {
if (err) {
logHelper(log, 'error', 'error heading object ' +
'from datastore', err, this._dataStoreName);
if (err.code === 'NotFound') {
const error = errors.ServiceUnavailable
.customizeDescription(
'Unexpected error from AWS: "NotFound". Data on AWS ' +
'may have been altered outside of CloudServer.'
);
return callback(error);
}
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
return callback();
});
}
get(objectGetInfo, range, reqUids, callback) {
const log = createLogger(reqUids);
const { key, dataStoreVersionId } = objectGetInfo;
const request = this._client.getObject({
Bucket: this._awsBucketName,
Key: key,
VersionId: dataStoreVersionId,
Range: range ? `bytes=${range[0]}-${range[1]}` : null,
}).on('success', response => {
log.trace('AWS GET request response headers',
{ responseHeaders: response.httpResponse.headers });
});
const stream = request.createReadStream().on('error', err => {
logHelper(log, 'error', 'error streaming data from AWS',
err, this._dataStoreName);
return callback(err);
});
return callback(null, stream);
}
delete(objectGetInfo, reqUids, callback) {
const { key, dataStoreVersionId, deleteVersion } = objectGetInfo;
const log = createLogger(reqUids);
const params = {
Bucket: this._awsBucketName,
Key: key,
};
if (deleteVersion) {
params.VersionId = dataStoreVersionId;
}
return this._client.deleteObject(params, err => {
if (err) {
logHelper(log, 'error', 'error deleting object from ' +
'datastore', err, this._dataStoreName);
if (err.code === 'NoSuchVersion') {
// data may have been deleted directly from the AWS backend
// don't want to retry the delete and errors are not
// sent back to client anyway, so no need to return err
return callback();
}
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
return callback();
});
}
healthcheck(location, callback) {
const awsResp = {};
this._client.headBucket({ Bucket: this._awsBucketName },
err => {
/* eslint-disable no-param-reassign */
if (err) {
awsResp[location] = { error: err, external: true };
return callback(null, awsResp);
}
return this._client.getBucketVersioning({
Bucket: this._awsBucketName },
(err, data) => {
if (err) {
awsResp[location] = { error: err, external: true };
} else if (!data.Status ||
data.Status === 'Suspended') {
awsResp[location] = {
versioningStatus: data.Status,
error: 'Versioning must be enabled',
external: true,
};
} else {
awsResp[location] = {
versioningStatus: data.Status,
message: 'Congrats! You own the bucket',
};
}
return callback(null, awsResp);
});
});
}
createMPU(key, metaHeaders, bucketName, websiteRedirectHeader, contentType,
cacheControl, contentDisposition, contentEncoding, log, callback) {
const metaHeadersTrimmed = {};
Object.keys(metaHeaders).forEach(header => {
if (header.startsWith('x-amz-meta-')) {
const headerKey = header.substring(11);
metaHeadersTrimmed[headerKey] = metaHeaders[header];
}
});
Object.assign(metaHeaders, metaHeadersTrimmed);
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucketName, key, this._bucketMatch);
const params = {
Bucket: awsBucket,
Key: awsKey,
WebsiteRedirectLocation: websiteRedirectHeader,
Metadata: metaHeaders,
ContentType: contentType,
CacheControl: cacheControl,
ContentDisposition: contentDisposition,
ContentEncoding: contentEncoding,
};
return this._client.createMultipartUpload(params, (err, mpuResObj) => {
if (err) {
logHelper(log, 'error', 'err from data backend',
err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
return callback(null, mpuResObj);
});
}
uploadPart(request, streamingV4Params, stream, size, key, uploadId,
partNumber, bucketName, log, callback) {
let hashedStream = stream;
if (request) {
const partStream = prepareStream(request, streamingV4Params,
log, callback);
hashedStream = new MD5Sum();
partStream.pipe(hashedStream);
}
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucketName, key, this._bucketMatch);
const params = { Bucket: awsBucket, Key: awsKey, UploadId: uploadId,
Body: hashedStream, ContentLength: size,
PartNumber: partNumber };
return this._client.uploadPart(params, (err, partResObj) => {
if (err) {
logHelper(log, 'error', 'err from data backend ' +
'on uploadPart', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
// Because we manually add quotes to ETag later, remove quotes here
const noQuotesETag =
partResObj.ETag.substring(1, partResObj.ETag.length - 1);
const dataRetrievalInfo = {
key: awsKey,
dataStoreType: 'aws_s3',
dataStoreName: this._dataStoreName,
dataStoreETag: noQuotesETag,
};
return callback(null, dataRetrievalInfo);
});
}
listParts(key, uploadId, bucketName, partNumberMarker, maxParts, log,
callback) {
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucketName, key, this._bucketMatch);
const params = { Bucket: awsBucket, Key: awsKey, UploadId: uploadId,
PartNumberMarker: partNumberMarker, MaxParts: maxParts };
return this._client.listParts(params, (err, partList) => {
if (err) {
logHelper(log, 'error', 'err from data backend on listPart',
err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
// build storedParts object to mimic Scality S3 backend returns
const storedParts = {};
storedParts.IsTruncated = partList.IsTruncated;
storedParts.Contents = [];
storedParts.Contents = partList.Parts.map(item => {
// We manually add quotes to ETag later, so remove quotes here
const noQuotesETag =
item.ETag.substring(1, item.ETag.length - 1);
return {
partNumber: item.PartNumber,
value: {
Size: item.Size,
ETag: noQuotesETag,
LastModified: item.LastModified,
},
};
});
return callback(null, storedParts);
});
}
/**
* completeMPU - complete multipart upload on AWS backend
* @param {object} jsonList - user-sent list of parts to include in
* final mpu object
* @param {object} mdInfo - object containing 3 keys: storedParts,
* mpuOverviewKey, and splitter
* @param {string} key - object key
* @param {string} uploadId - multipart upload id string
* @param {string} bucketName - name of bucket
* @param {RequestLogger} log - logger instance
* @param {function} callback - callback function
* @return {(Error|object)} - return Error if complete MPU fails, otherwise
* object containing completed object key, eTag, and contentLength
*/
completeMPU(jsonList, mdInfo, key, uploadId, bucketName, log, callback) {
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucketName, key, this._bucketMatch);
const mpuError = {
InvalidPart: true,
InvalidPartOrder: true,
EntityTooSmall: true,
};
const partArray = [];
const partList = jsonList.Part;
partList.forEach(partObj => {
const partParams = { PartNumber: partObj.PartNumber[0],
ETag: partObj.ETag[0] };
partArray.push(partParams);
});
const mpuParams = {
Bucket: awsBucket, Key: awsKey, UploadId: uploadId,
MultipartUpload: {
Parts: partArray,
},
};
const completeObjData = { key: awsKey };
return this._client.completeMultipartUpload(mpuParams,
(err, completeMpuRes) => {
if (err) {
if (mpuError[err.code]) {
logHelper(log, 'trace', 'err from data backend on ' +
'completeMPU', err, this._dataStoreName);
return callback(errors[err.code]);
}
logHelper(log, 'error', 'err from data backend on ' +
'completeMPU', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
if (!completeMpuRes.VersionId) {
logHelper(log, 'error', 'missing version id for data ' +
'backend object', missingVerIdInternalError,
this._dataStoreName);
return callback(missingVerIdInternalError);
}
// need to get content length of new object to store
// in our metadata
return this._client.headObject({ Bucket: awsBucket, Key: awsKey },
(err, objHeaders) => {
if (err) {
logHelper(log, 'trace', 'err from data backend on ' +
'headObject', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
// remove quotes from eTag because they're added later
completeObjData.eTag = completeMpuRes.ETag
.substring(1, completeMpuRes.ETag.length - 1);
completeObjData.dataStoreVersionId = completeMpuRes.VersionId;
completeObjData.contentLength = objHeaders.ContentLength;
return callback(null, completeObjData);
});
});
}
abortMPU(key, uploadId, bucketName, log, callback) {
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucketName, key, this._bucketMatch);
const abortParams = {
Bucket: awsBucket, Key: awsKey, UploadId: uploadId,
};
return this._client.abortMultipartUpload(abortParams, err => {
if (err) {
logHelper(log, 'error', 'There was an error aborting ' +
'the MPU on AWS S3. You should abort directly on AWS S3 ' +
'using the same uploadId.', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
return callback();
});
}
objectPutTagging(key, bucket, objectMD, log, callback) {
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucket, key, this._bucketMatch);
const dataStoreVersionId = objectMD.location[0].dataStoreVersionId;
const tagParams = {
Bucket: awsBucket,
Key: awsKey,
VersionId: dataStoreVersionId,
};
const keyArray = Object.keys(objectMD.tags);
tagParams.Tagging = {};
tagParams.Tagging.TagSet = keyArray.map(key => {
const value = objectMD.tags[key];
return { Key: key, Value: value };
});
return this._client.putObjectTagging(tagParams, err => {
if (err) {
logHelper(log, 'error', 'error from data backend on ' +
'putObjectTagging', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
return callback();
});
}
objectDeleteTagging(key, bucket, objectMD, log, callback) {
const awsBucket = this._awsBucketName;
const awsKey = this._createAwsKey(bucket, key, this._bucketMatch);
const dataStoreVersionId = objectMD.location[0].dataStoreVersionId;
const tagParams = {
Bucket: awsBucket,
Key: awsKey,
VersionId: dataStoreVersionId,
};
return this._client.deleteObjectTagging(tagParams, err => {
if (err) {
logHelper(log, 'error', 'error from data backend on ' +
'deleteObjectTagging', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
return callback();
});
}
copyObject(request, destLocationConstraintName, sourceKey,
sourceLocationConstraintName, storeMetadataParams, log, callback) {
const destBucketName = request.bucketName;
const destObjectKey = request.objectKey;
const destAwsKey = this._createAwsKey(destBucketName, destObjectKey,
this._bucketMatch);
const sourceAwsBucketName =
config.getAwsBucketName(sourceLocationConstraintName);
const metadataDirective = request.headers['x-amz-metadata-directive'];
const metaHeaders = trimXMetaPrefix(getMetaHeaders(request.headers));
const awsParams = {
Bucket: this._awsBucketName,
Key: destAwsKey,
CopySource: `${sourceAwsBucketName}/${sourceKey}`,
Metadata: metaHeaders,
MetadataDirective: metadataDirective,
};
if (destLocationConstraintName &&
config.isAWSServerSideEncrytion(destLocationConstraintName)) {
awsParams.ServerSideEncryption = 'AES256';
}
this._client.copyObject(awsParams, (err, copyResult) => {
if (err) {
if (err.code === 'AccessDenied') {
logHelper(log, 'error', 'Unable to access ' +
`${sourceAwsBucketName} AWS bucket`, err,
this._dataStoreName);
return callback(errors.AccessDenied
.customizeDescription('Error: Unable to access ' +
`${sourceAwsBucketName} AWS bucket`)
);
}
logHelper(log, 'error', 'error from data backend on ' +
'copyObject', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
if (!copyResult.VersionId) {
logHelper(log, 'error', 'missing version id for data ' +
'backend object', missingVerIdInternalError,
this._dataStoreName);
return callback(missingVerIdInternalError);
}
return callback(null, destAwsKey, copyResult.VersionId);
});
}
uploadPartCopy(request, awsSourceKey, sourceLocationConstraintName,
log, callback) {
const destBucketName = request.bucketName;
const destObjectKey = request.objectKey;
const destAwsKey = this._createAwsKey(destBucketName, destObjectKey,
this._bucketMatch);
const sourceAwsBucketName =
config.getAwsBucketName(sourceLocationConstraintName);
const uploadId = request.query.uploadId;
const partNumber = request.query.partNumber;
const copySourceRange = request.headers['x-amz-copy-source-range'];
const params = {
Bucket: this._awsBucketName,
CopySource: `${sourceAwsBucketName}/${awsSourceKey}`,
CopySourceRange: copySourceRange,
Key: destAwsKey,
PartNumber: partNumber,
UploadId: uploadId,
};
return this._client.uploadPartCopy(params, (err, res) => {
if (err) {
if (err.code === 'AccessDenied') {
logHelper(log, 'error', 'Unable to access ' +
`${sourceAwsBucketName} AWS bucket`, err,
this._dataStoreName);
return callback(errors.AccessDenied
.customizeDescription('Error: Unable to access ' +
`${sourceAwsBucketName} AWS bucket`)
);
}
logHelper(log, 'error', 'error from data backend on ' +
'uploadPartCopy', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
const eTag = removeQuotes(res.CopyPartResult.ETag);
return callback(null, eTag);
});
}
}
module.exports = AwsClient;

434
backends/data/external/AzureClient.js vendored Normal file
View File

@ -0,0 +1,434 @@
const url = require('url');
const { errors, s3middleware } = require('arsenal');
const azure = require('azure-storage');
const createLogger = require('../multipleBackendLogger');
const { logHelper, translateAzureMetaHeaders } = require('./utils');
const { config } = require('../../Config');
const { validateAndFilterMpuParts } =
require('../../api/apiUtils/object/processMpuParts');
const constants = require('../../../constants');
const metadata = require('../../metadata/wrapper');
const azureMpuUtils = s3middleware.azureHelper.mpuUtils;
class AzureClient {
constructor(config) {
this._azureStorageEndpoint = config.azureStorageEndpoint;
this._azureStorageCredentials = config.azureStorageCredentials;
this._azureContainerName = config.azureContainerName;
this._client = azure.createBlobService(
this._azureStorageCredentials.storageAccountName,
this._azureStorageCredentials.storageAccessKey,
this._azureStorageEndpoint);
this._dataStoreName = config.dataStoreName;
this._bucketMatch = config.bucketMatch;
if (config.proxy) {
const parsedUrl = url.parse(config.proxy);
if (!parsedUrl.port) {
parsedUrl.port = 80;
}
this._client.setProxy(parsedUrl);
}
}
_errorWrapper(s3Method, azureMethod, args, log, cb) {
if (log) {
log.info(`calling azure ${azureMethod}`);
}
try {
this._client[azureMethod].apply(this._client, args);
} catch (err) {
const error = errors.ServiceUnavailable;
if (log) {
log.error('error thrown by Azure Storage Client Library',
{ error: err.message, stack: err.stack, s3Method,
azureMethod, dataStoreName: this._dataStoreName });
}
cb(error.customizeDescription('Error from Azure ' +
`method: ${azureMethod} on ${s3Method} S3 call: ` +
`${err.message}`));
}
}
_createAzureKey(requestBucketName, requestObjectKey,
bucketMatch) {
if (bucketMatch) {
return requestObjectKey;
}
return `${requestBucketName}/${requestObjectKey}`;
}
_getMetaHeaders(objectMD) {
const metaHeaders = {};
Object.keys(objectMD).forEach(mdKey => {
const isMetaHeader = mdKey.startsWith('x-amz-meta-');
if (isMetaHeader) {
metaHeaders[mdKey] = objectMD[mdKey];
}
});
return translateAzureMetaHeaders(metaHeaders);
}
// Before putting or deleting object on Azure, check if MPU exists with
// same key name. If it does, do not allow put or delete because Azure
// will delete all blocks with same key name
protectAzureBlocks(bucketName, objectKey, dataStoreName, log, cb) {
const mpuBucketName = `${constants.mpuBucketPrefix}${bucketName}`;
const splitter = constants.splitter;
const listingParams = {
prefix: `overview${splitter}${objectKey}`,
listingType: 'MPU',
splitter,
maxKeys: 1,
};
return metadata.listMultipartUploads(mpuBucketName, listingParams,
log, (err, mpuList) => {
if (err && !err.NoSuchBucket) {
log.error('Error listing MPUs for Azure delete',
{ error: err, dataStoreName });
return cb(errors.ServiceUnavailable);
}
if (mpuList && mpuList.Uploads && mpuList.Uploads.length > 0) {
const error = errors.MPUinProgress;
log.error('Error: cannot put/delete object to Azure with ' +
'same key name as ongoing MPU on Azure',
{ error, dataStoreName });
return cb(error);
}
// If listMultipartUploads returns a NoSuchBucket error or the
// mpu list is empty, there are no conflicting MPUs, so continue
return cb();
});
}
put(stream, size, keyContext, reqUids, callback) {
const log = createLogger(reqUids);
// before blob is put, make sure there is no ongoing MPU with same key
this.protectAzureBlocks(keyContext.bucketName,
keyContext.objectKey, this._dataStoreName, log, err => {
// if error returned, there is ongoing MPU, so do not put
if (err) {
return callback(err.customizeDescription(
`Error putting object to Azure: ${err.message}`));
}
const azureKey = this._createAzureKey(keyContext.bucketName,
keyContext.objectKey, this._bucketMatch);
const options = {
metadata: translateAzureMetaHeaders(keyContext.metaHeaders,
keyContext.tagging),
contentSettings: {
contentType: keyContext.contentType || undefined,
cacheControl: keyContext.cacheControl || undefined,
contentDisposition: keyContext.contentDisposition ||
undefined,
contentEncoding: keyContext.contentEncoding || undefined,
},
};
if (size === 0) {
return this._errorWrapper('put', 'createBlockBlobFromText',
[this._azureContainerName, azureKey, '', options,
err => {
if (err) {
logHelper(log, 'error', 'err from Azure PUT data ' +
'backend', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`Azure: ${err.message}`));
}
return callback(null, azureKey);
}], log, callback);
}
return this._errorWrapper('put', 'createBlockBlobFromStream',
[this._azureContainerName, azureKey, stream, size, options,
err => {
if (err) {
logHelper(log, 'error', 'err from Azure PUT data ' +
'backend', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`Azure: ${err.message}`));
}
return callback(null, azureKey);
}], log, callback);
});
}
head(objectGetInfo, reqUids, callback) {
const log = createLogger(reqUids);
const { key, azureStreamingOptions } = objectGetInfo;
return this._errorWrapper('head', 'getBlobProperties',
[this._azureContainerName, key, azureStreamingOptions,
err => {
if (err) {
logHelper(log, 'error', 'err from Azure HEAD data backend',
err, this._dataStoreName);
if (err.code === 'NotFound') {
const error = errors.ServiceUnavailable
.customizeDescription(
'Unexpected error from Azure: "NotFound". Data ' +
'on Azure may have been altered outside of ' +
'CloudServer.');
return callback(error);
}
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`Azure: ${err.message}`));
}
return callback();
}], log, callback);
}
get(objectGetInfo, range, reqUids, callback) {
const log = createLogger(reqUids);
// for backwards compatibility
const { key, response, azureStreamingOptions } = objectGetInfo;
let streamingOptions;
if (azureStreamingOptions) {
// option coming from api.get()
streamingOptions = azureStreamingOptions;
} else if (range) {
// option coming from multipleBackend.upload()
const rangeStart = range[0] ? range[0].toString() : undefined;
const rangeEnd = range[1] ? range[1].toString() : undefined;
streamingOptions = { rangeStart, rangeEnd };
}
this._errorWrapper('get', 'getBlobToStream',
[this._azureContainerName, key, response, streamingOptions,
err => {
if (err) {
logHelper(log, 'error', 'err from Azure GET data backend',
err, this._dataStoreName);
return callback(errors.ServiceUnavailable);
}
return callback(null, response);
}], log, callback);
}
delete(objectGetInfo, reqUids, callback) {
const log = createLogger(reqUids);
// for backwards compatibility
const key = typeof objectGetInfo === 'string' ? objectGetInfo :
objectGetInfo.key;
return this._errorWrapper('delete', 'deleteBlobIfExists',
[this._azureContainerName, key,
err => {
if (err) {
const log = createLogger(reqUids);
logHelper(log, 'error', 'error deleting object from ' +
'Azure datastore', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`Azure: ${err.message}`));
}
return callback();
}], log, callback);
}
healthcheck(location, callback, flightCheckOnStartUp) {
const azureResp = {};
const healthCheckAction = flightCheckOnStartUp ?
'createContainerIfNotExists' : 'doesContainerExist';
this._errorWrapper('checkAzureHealth', healthCheckAction,
[this._azureContainerName, err => {
/* eslint-disable no-param-reassign */
if (err) {
azureResp[location] = { error: err.message,
external: true };
return callback(null, azureResp);
}
azureResp[location] = {
message:
'Congrats! You can access the Azure storage account',
};
return callback(null, azureResp);
}], null, callback);
}
uploadPart(request, streamingV4Params, partStream, size, key, uploadId,
partNumber, bucket, log, callback) {
const azureKey = this._createAzureKey(bucket, key, this._bucketMatch);
const params = { bucketName: this._azureContainerName,
partNumber, size, objectKey: azureKey, uploadId };
const stream = request || partStream;
if (request && request.headers && request.headers['content-md5']) {
params.contentMD5 = request.headers['content-md5'];
}
const dataRetrievalInfo = {
key: azureKey,
partNumber,
dataStoreName: this._dataStoreName,
dataStoreType: 'azure',
numberSubParts: azureMpuUtils.getSubPartInfo(size)
.expectedNumberSubParts,
};
if (size === 0) {
log.debug('0-byte part does not store data',
{ method: 'uploadPart' });
dataRetrievalInfo.dataStoreETag = azureMpuUtils.zeroByteETag;
dataRetrievalInfo.numberSubParts = 0;
return callback(null, dataRetrievalInfo);
}
if (size <= azureMpuUtils.maxSubPartSize) {
const errorWrapperFn = this._errorWrapper.bind(this);
return azureMpuUtils.putSinglePart(errorWrapperFn,
stream, params, this._dataStoreName, log, (err, dataStoreETag) => {
if (err) {
return callback(err);
}
dataRetrievalInfo.dataStoreETag = dataStoreETag;
return callback(null, dataRetrievalInfo);
});
}
const errorWrapperFn = this._errorWrapper.bind(this);
return azureMpuUtils.putSubParts(errorWrapperFn, stream,
params, this._dataStoreName, log, (err, dataStoreETag) => {
if (err) {
return callback(err);
}
dataRetrievalInfo.dataStoreETag = dataStoreETag;
return callback(null, dataRetrievalInfo);
});
}
completeMPU(jsonList, mdInfo, key, uploadId, bucket, metaHeaders,
contentSettings, log, callback) {
const azureKey = this._createAzureKey(bucket, key, this._bucketMatch);
const commitList = {
UncommittedBlocks: jsonList.uncommittedBlocks || [],
};
let filteredPartsObj;
if (!jsonList.uncommittedBlocks) {
const { storedParts, mpuOverviewKey, splitter } = mdInfo;
filteredPartsObj = validateAndFilterMpuParts(storedParts, jsonList,
mpuOverviewKey, splitter, log);
filteredPartsObj.partList.forEach(part => {
// part.locations is always array of 1, which contains data info
const subPartIds =
azureMpuUtils.getSubPartIds(part.locations[0], uploadId);
commitList.UncommittedBlocks.push(...subPartIds);
});
}
const options = {
contentSettings,
metadata: metaHeaders ? translateAzureMetaHeaders(metaHeaders) :
undefined,
};
this._errorWrapper('completeMPU', 'commitBlocks',
[this._azureContainerName, azureKey, commitList, options,
err => {
if (err) {
logHelper(log, 'error', 'err completing MPU on Azure ' +
'datastore', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`Azure: ${err.message}`));
}
const completeObjData = {
key: azureKey,
filteredPartsObj,
};
return callback(null, completeObjData);
}], log, callback);
}
objectPutTagging(key, bucket, objectMD, log, callback) {
const azureKey = this._createAzureKey(bucket, key, this._bucketMatch);
const azureMD = this._getMetaHeaders(objectMD);
azureMD.tags = JSON.stringify(objectMD.tags);
this._errorWrapper('objectPutTagging', 'setBlobMetadata',
[this._azureContainerName, azureKey, azureMD,
err => {
if (err) {
logHelper(log, 'error', 'err putting object tags to ' +
'Azure backend', err, this._dataStoreName);
return callback(errors.ServiceUnavailable);
}
return callback();
}], log, callback);
}
objectDeleteTagging(key, bucket, objectMD, log, callback) {
const azureKey = this._createAzureKey(bucket, key, this._bucketMatch);
const azureMD = this._getMetaHeaders(objectMD);
this._errorWrapper('objectDeleteTagging', 'setBlobMetadata',
[this._azureContainerName, azureKey, azureMD,
err => {
if (err) {
logHelper(log, 'error', 'err putting object tags to ' +
'Azure backend', err, this._dataStoreName);
return callback(errors.ServiceUnavailable);
}
return callback();
}], log, callback);
}
copyObject(request, destLocationConstraintName, sourceKey,
sourceLocationConstraintName, storeMetadataParams, log, callback) {
const destContainerName = request.bucketName;
const destObjectKey = request.objectKey;
const destAzureKey = this._createAzureKey(destContainerName,
destObjectKey, this._bucketMatch);
const sourceContainerName =
config.locationConstraints[sourceLocationConstraintName]
.details.azureContainerName;
let options;
if (storeMetadataParams.metaHeaders) {
options = { metadata:
translateAzureMetaHeaders(storeMetadataParams.metaHeaders) };
}
this._errorWrapper('copyObject', 'startCopyBlob',
[`${this._azureStorageEndpoint}` +
`${sourceContainerName}/${sourceKey}`,
this._azureContainerName, destAzureKey, options,
(err, res) => {
if (err) {
if (err.code === 'CannotVerifyCopySource') {
logHelper(log, 'error', 'Unable to access ' +
`${sourceContainerName} Azure Container`, err,
this._dataStoreName);
return callback(errors.AccessDenied
.customizeDescription('Error: Unable to access ' +
`${sourceContainerName} Azure Container`)
);
}
logHelper(log, 'error', 'error from data backend on ' +
'copyObject', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS: ${err.message}`)
);
}
if (res.copy.status === 'pending') {
logHelper(log, 'error', 'Azure copy status is pending',
err, this._dataStoreName);
const copyId = res.copy.id;
this._client.abortCopyBlob(this._azureContainerName,
destAzureKey, copyId, err => {
if (err) {
logHelper(log, 'error', 'error from data backend ' +
'on abortCopyBlob', err, this._dataStoreName);
return callback(errors.ServiceUnavailable
.customizeDescription('Error returned from ' +
`AWS on abortCopyBlob: ${err.message}`)
);
}
return callback(errors.InvalidObjectState
.customizeDescription('Error: Azure copy status was ' +
'pending. It has been aborted successfully')
);
});
}
return callback(null, destAzureKey);
}], log, callback);
}
}
module.exports = AzureClient;

246
backends/data/external/GCP/GcpService.js vendored Normal file
View File

@ -0,0 +1,246 @@
const googleAuth = require('google-auto-auth');
const async = require('async');
const AWS = require('aws-sdk');
const { errors } = require('arsenal');
const Service = AWS.Service;
const GcpSigner = require('./GcpSigner');
function genAuth(authParams, callback) {
async.tryEach([
function authKeyFile(next) {
const authOptions = {
scopes: authParams.scopes,
keyFilename: authParams.keyFilename,
};
const auth = googleAuth(authOptions);
auth.getToken(err => next(err, auth));
},
function authCredentials(next) {
const authOptions = {
scopes: authParams.scopes,
credentials: authParams.credentials,
};
const auth = googleAuth(authOptions);
auth.getToken(err => next(err, auth));
},
], (err, result) => callback(err, result));
}
AWS.apiLoader.services.gcp = {};
const GCP = Service.defineService('gcp', ['2017-11-01'], {
_jsonAuth: null,
_authParams: null,
getToken(callback) {
if (this._jsonAuth) {
return this._jsonAuth.getToken(callback);
}
if (!this._authParams && this.config.authParams &&
typeof this.config.authParams === 'object') {
this._authParams = this.config.authParams;
}
return genAuth(this._authParams, (err, auth) => {
if (!err) {
this._jsonAuth = auth;
return this._jsonAuth.getToken(callback);
}
// should never happen, but if all preconditions fails
// can't generate tokens
return callback(errors.InternalError.customizeDescription(
'Unable to create a google authorizer'));
});
},
getSignerClass() {
return GcpSigner;
},
validateService() {
if (!this.config.region) {
this.config.region = 'us-east-1';
}
},
upload(params, options, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: upload not implemented'));
},
// Service API
listBuckets(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: listBuckets not implemented'));
},
// Bucket APIs
getBucketLocation(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: getBucketLocation not implemented'));
},
deleteBucket(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: deleteBucket not implemented'));
},
headBucket(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: headBucket not implemented'));
},
listObjects(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: listObjects not implemented'));
},
listObjectVersions(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: listObjecVersions not implemented'));
},
putBucket(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putBucket not implemented'));
},
getBucketAcl(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: getBucketAcl not implemented'));
},
putBucketAcl(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putBucketAcl not implemented'));
},
putBucketWebsite(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putBucketWebsite not implemented'));
},
getBucketWebsite(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: getBucketWebsite not implemented'));
},
deleteBucketWebsite(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: deleteBucketWebsite not implemented'));
},
putBucketCors(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putBucketCors not implemented'));
},
getBucketCors(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: getBucketCors not implemented'));
},
deleteBucketCors(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: deleteBucketCors not implemented'));
},
// Object APIs
headObject(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: headObject not implemented'));
},
putObject(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putObject not implemented'));
},
getObject(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: getObject not implemented'));
},
deleteObject(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: deleteObject not implemented'));
},
deleteObjects(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: deleteObjects not implemented'));
},
copyObject(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: copyObject not implemented'));
},
putObjectTagging(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putObjectTagging not implemented'));
},
deleteObjectTagging(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: deleteObjectTagging not implemented'));
},
putObjectAcl(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: putObjectAcl not implemented'));
},
getObjectAcl(params, callback) {
return callback(errors.NotImplemented
.customizeDescription('GCP: getObjectAcl not implemented'));
},
// Multipart upload
abortMultipartUpload(params, callback) {
return callback(errors.NotImplemented
.customizeDescription(
'GCP: abortMultipartUpload not implemented'));
},
createMultipartUpload(params, callback) {
return callback(errors.NotImplemented
.customizeDescription(
'GCP: createMultipartUpload not implemented'));
},
completeMultipartUpload(params, callback) {
return callback(errors.NotImplemented
.customizeDescription(
'GCP: completeMultipartUpload not implemented'));
},
uploadPart(params, callback) {
return callback(errors.NotImplemented
.customizeDescription(
'GCP: uploadPart not implemented'));
},
uploadPartCopy(params, callback) {
return callback(errors.NotImplemented
.customizeDescription(
'GCP: uploadPartCopy not implemented'));
},
listParts(params, callback) {
return callback(errors.NotImplemented
.customizeDescription(
'GCP: listParts not implemented'));
},
});
Object.defineProperty(AWS.apiLoader.services.gcp, '2017-11-01', {
get: function get() {
const model = require('./gcp-2017-11-01.api.json');
return model;
},
enumerable: true,
configurable: true,
});
module.exports = GCP;

48
backends/data/external/GCP/GcpSigner.js vendored Normal file
View File

@ -0,0 +1,48 @@
const url = require('url');
const qs = require('querystring');
const AWS = require('aws-sdk');
const werelogs = require('werelogs');
const { constructStringToSignV2 } = require('arsenal').auth.client;
const logger = new werelogs.Logger('GcpSigner');
function genQueryObject(uri) {
const queryString = url.parse(uri).query;
return qs.parse(queryString);
}
const GcpSigner = AWS.util.inherit(AWS.Signers.RequestSigner, {
constructor: function GcpSigner(request) {
AWS.Signers.RequestSigner.call(this, request);
},
addAuthorization: function addAuthorization(credentials, date) {
if (!this.request.headers['presigned-expires']) {
this.request.headers['x-goog-date'] = AWS.util.date.rfc822(date);
}
const signature =
this.sign(credentials.secretAccessKey, this.stringToSign());
const auth = `GOOG1 ${credentials.accessKeyId}: ${signature}`;
this.request.headers.Authorization = auth;
},
stringToSign: function stringToSign() {
const requestObject = {
url: this.request.path,
method: this.request.method,
host: this.request.endpoint.host,
headers: this.request.headers,
query: genQueryObject(this.request.path) || {},
};
const data = Object.assign({}, this.request.headers);
return constructStringToSignV2(requestObject, data, logger, 'GCP');
},
sign: function sign(secret, string) {
return AWS.util.crypto.hmac(secret, string, 'base64', 'sha1');
},
});
module.exports = GcpSigner;

View File

@ -0,0 +1 @@
module.exports = {};

View File

@ -0,0 +1,17 @@
{
"version": "1.0",
"metadata": {
"apiVersion": "2017-11-01",
"checksumFormat": "md5",
"endpointPrefix": "s3",
"globalEndpoint": "storage.googleapi.com",
"protocol": "rest-xml",
"serviceAbbreviation": "GCP",
"serviceFullName": "Google Cloud Storage",
"signatureVersion": "s3",
"timestampFormat": "rfc822",
"uid": "gcp-2017-11-01"
},
"operations": {},
"shapes": {}
}

5
backends/data/external/GCP/index.js vendored Normal file
View File

@ -0,0 +1,5 @@
module.exports = {
GCP: require('./GcpService'),
GcpSigner: require('./GcpSigner'),
GcpUtils: require('./GcpUtils'),
};

39
backends/data/external/GcpClient.js vendored Normal file
View File

@ -0,0 +1,39 @@
const { GCP } = require('./GCP');
const AwsClient = require('./AwsClient');
/**
* Class representing a Google Cloud Storage backend object
* @extends AwsClient
*/
class GcpClient extends AwsClient {
/**
* constructor - creates a Gcp backend client object (inherits )
* @param {object} config - configuration object for Gcp Backend up
* @param {object} config.s3params - S3 configuration
* @param {string} config.bucketName - GCP bucket name
* @param {string} config.mpuBucket - GCP mpu bucket name
* @param {string} config.overflowBucket - GCP overflow bucket name
* @param {boolean} config.bucketMatch - bucket match flag
* @param {object} config.authParams - GCP service credentials
* @param {string} config.dataStoreName - locationConstraint name
* @param {booblean} config.serverSideEncryption - server side encryption
* flag
* @return {object} - returns a GcpClient object
*/
constructor(config) {
super(config);
this.clientType = 'gcp';
this._gcpBucketName = config.bucketName;
this._mpuBucketName = config.mpuBucket;
this._overflowBucketname = config.overflowBucket;
this._gcpParams = Object.assign(this._s3Params, {
mainBucket: this._gcpBucketName,
mpuBucket: this._mpuBucketName,
overflowBucket: this._overflowBucketname,
authParams: config.authParams,
});
this._client = new GCP(this._gcpParams);
}
}
module.exports = GcpClient;

147
backends/data/external/utils.js vendored Normal file
View File

@ -0,0 +1,147 @@
const async = require('async');
const constants = require('../../../constants');
const { config } = require('../../../lib/Config');
const awsHealth = {
response: undefined,
time: 0,
};
const azureHealth = {
response: undefined,
time: 0,
};
const utils = {
logHelper(log, level, description, error, dataStoreName) {
log[level](description, { error: error.message,
errorName: error.name, dataStoreName });
},
// take off the 'x-amz-meta-'
trimXMetaPrefix(obj) {
const newObj = {};
Object.keys(obj).forEach(key => {
const newKey = key.substring(11);
newObj[newKey] = obj[key];
});
return newObj;
},
removeQuotes(word) {
return word.slice(1, -1);
},
skipMpuPartProcessing(completeObjData) {
const backendType = completeObjData.dataStoreType;
if (constants.mpuMDStoredExternallyBackend[backendType]) {
return true;
}
return false;
},
/**
* checkAzureBackendMatch - checks that the external backend location for
* two data objects is the same and is Azure
* @param {array} objectDataOne - data of first object to compare
* @param {object} objectDataTwo - data of second object to compare
* @return {boolean} - true if both data backends are Azure, false if not
*/
checkAzureBackendMatch(objectDataOne, objectDataTwo) {
if (objectDataOne.dataStoreType === 'azure' &&
objectDataTwo.dataStoreType === 'azure') {
return true;
}
return false;
},
/**
* externalBackendCopy - Server side copy should only be allowed:
* 1) if source object and destination object are both on aws or both
* on azure.
* 2) if azure to azure, must be the same storage account since Azure
* copy outside of an account is async
* 3) if the source bucket is not an encrypted bucket and the
* destination bucket is not an encrypted bucket (unless the copy
* is all within the same bucket).
* @param {string} locationConstraintSrc - location constraint of the source
* @param {string} locationConstraintDest - location constraint of the
* destination
* @param {object} sourceBucketMD - metadata of the source bucket
* @param {object} destBucketMD - metadata of the destination bucket
* @return {boolean} - true if copying object from one
* externalbackend to the same external backend and for Azure if it is the
* same account since Azure copy outside of an account is async
*/
externalBackendCopy(locationConstraintSrc, locationConstraintDest,
sourceBucketMD, destBucketMD) {
const sourceBucketName = sourceBucketMD.getName();
const destBucketName = destBucketMD.getName();
const isSameBucket = sourceBucketName === destBucketName;
const bucketsNotEncrypted = destBucketMD.getServerSideEncryption()
=== sourceBucketMD.getServerSideEncryption() &&
sourceBucketMD.getServerSideEncryption() === null;
const sourceLocationConstraintType =
config.getLocationConstraintType(locationConstraintSrc);
const locationTypeMatch =
config.getLocationConstraintType(locationConstraintSrc) ===
config.getLocationConstraintType(locationConstraintDest);
return locationTypeMatch && (isSameBucket || bucketsNotEncrypted) &&
(sourceLocationConstraintType === 'aws_s3' ||
(sourceLocationConstraintType === 'azure' &&
config.isSameAzureAccount(locationConstraintSrc,
locationConstraintDest)));
},
checkExternalBackend(clients, locations, type, flightCheckOnStartUp,
externalBackendHealthCheckInterval, cb) {
const checkStatus = type === 'aws_s3' ? awsHealth : azureHealth;
if (locations.length === 0) {
return process.nextTick(cb, null, []);
}
if (!flightCheckOnStartUp && checkStatus.response &&
Date.now() - checkStatus.time
< externalBackendHealthCheckInterval) {
return process.nextTick(cb, null, checkStatus.response);
}
let locationsToCheck;
if (flightCheckOnStartUp) {
// check all locations if flight check on start up
locationsToCheck = locations;
} else {
const randomLocation = locations[Math.floor(Math.random() *
locations.length)];
locationsToCheck = [randomLocation];
}
return async.mapLimit(locationsToCheck, 5, (location, next) => {
const client = clients[location];
client.healthcheck(location, next, flightCheckOnStartUp);
}, (err, results) => {
if (err) {
return cb(err);
}
if (!flightCheckOnStartUp) {
checkStatus.response = results;
checkStatus.time = Date.now();
}
return cb(null, results);
});
},
translateAzureMetaHeaders(metaHeaders, tags) {
const translatedMetaHeaders = {};
if (tags) {
// tags are passed as string of format 'key1=value1&key2=value2'
const tagObj = {};
const tagArr = tags.split('&');
tagArr.forEach(keypair => {
const equalIndex = keypair.indexOf('=');
const key = keypair.substring(0, equalIndex);
tagObj[key] = keypair.substring(equalIndex + 1);
});
translatedMetaHeaders.tags = JSON.stringify(tagObj);
}
Object.keys(metaHeaders).forEach(headerName => {
const translated = headerName.substring(11).replace(/-/g, '_');
translatedMetaHeaders[translated] = metaHeaders[headerName];
});
return translatedMetaHeaders;
},
};
module.exports = utils;

View File

@ -0,0 +1,37 @@
const arsenal = require('arsenal');
const { config } = require('../../Config');
class DataFileInterface {
constructor() {
const { host, port } = config.dataClient;
this.restClient = new arsenal.network.rest.RESTClient(
{ host, port });
}
put(stream, size, keyContext, reqUids, callback) {
// ignore keyContext
this.restClient.put(stream, size, reqUids, callback);
}
get(objectGetInfo, range, reqUids, callback) {
const key = objectGetInfo.key ? objectGetInfo.key : objectGetInfo;
this.restClient.get(key, range, reqUids, callback);
}
delete(objectGetInfo, reqUids, callback) {
const key = objectGetInfo.key ? objectGetInfo.key : objectGetInfo;
this.restClient.delete(key, reqUids, callback);
}
getDiskUsage(reqUids, callback) {
this.restClient.getAction('diskUsage', reqUids, (err, val) => {
if (err) {
return callback(err);
}
return callback(null, JSON.parse(val));
});
}
}
module.exports = DataFileInterface;

View File

@ -0,0 +1,93 @@
const stream = require('stream');
const { errors } = require('arsenal');
const werelogs = require('werelogs');
const logger = new werelogs.Logger('MemDataBackend');
function createLogger(reqUids) {
return reqUids ?
logger.newRequestLoggerFromSerializedUids(reqUids) :
logger.newRequestLogger();
}
const ds = [];
let count = 1; // keys are assessed with if (!key)
function resetCount() {
count = 1;
}
const backend = {
put: function putMem(request, size, keyContext, reqUids, callback) {
const log = createLogger(reqUids);
const value = Buffer.alloc(size);
let cursor = 0;
let exceeded = false;
request.on('data', data => {
if (cursor + data.length > size) {
exceeded = true;
}
if (!exceeded) {
data.copy(value, cursor);
}
cursor += data.length;
})
.on('end', () => {
if (exceeded) {
log.error('data stream exceed announced size',
{ size, overflow: cursor });
callback(errors.InternalError);
} else {
ds[count] = { value, keyContext };
callback(null, count++);
}
});
},
get: function getMem(objectGetInfo, range, reqUids, callback) {
const key = objectGetInfo.key ? objectGetInfo.key : objectGetInfo;
process.nextTick(() => {
if (!ds[key]) { return callback(errors.NoSuchKey); }
const storedBuffer = ds[key].value;
// If a range was sent, use the start from the range.
// Otherwise, start at 0
let start = range ? range[0] : 0;
// If a range was sent, use the end from the range.
// End of range should be included so +1
// Otherwise, get the full length
const end = range ? range[1] + 1 : storedBuffer.length;
const chunkSize = 64 * 1024; // 64KB
const val = new stream.Readable({
read: function read() {
// sets this._read under the hood
// push data onto the read queue, passing null
// will signal the end of the stream (EOF)
while (start < end) {
const finish =
Math.min(start + chunkSize, end);
this.push(storedBuffer.slice(start, finish));
start += chunkSize;
}
if (start >= end) {
this.push(null);
}
},
});
return callback(null, val);
});
},
delete: function delMem(objectGetInfo, reqUids, callback) {
const key = objectGetInfo.key ? objectGetInfo.key : objectGetInfo;
process.nextTick(() => {
delete ds[key];
return callback(null);
});
},
};
module.exports = {
backend,
ds,
resetCount,
};

View File

@ -0,0 +1,126 @@
const http = require('http');
const https = require('https');
const AWS = require('aws-sdk');
const Sproxy = require('sproxydclient');
const DataFileBackend = require('./file/backend');
const inMemory = require('./in_memory/backend').backend;
const AwsClient = require('./external/AwsClient');
const GcpClient = require('./external/GcpClient');
const AzureClient = require('./external/AzureClient');
const { config } = require('../Config');
const proxyAddress = 'http://localhost:3128';
function parseLC() {
const clients = {};
Object.keys(config.locationConstraints).forEach(location => {
const locationObj = config.locationConstraints[location];
if (locationObj.type === 'mem') {
clients[location] = inMemory;
}
if (locationObj.type === 'file') {
clients[location] = new DataFileBackend();
}
if (locationObj.type === 'scality'
&& locationObj.details.connector.sproxyd) {
clients[location] = new Sproxy({
bootstrap: locationObj.details.connector
.sproxyd.bootstrap,
// Might be undefined which is ok since there is a default
// set in sproxydclient if chordCos is undefined
chordCos: locationObj.details.connector.sproxyd.chordCos,
// Might also be undefined, but there is a default path set
// in sproxydclient as well
path: locationObj.details.connector.sproxyd.path,
// enable immutable optim for all objects
immutable: true,
});
clients[location].clientType = 'scality';
}
if (locationObj.type === 'aws_s3' || locationObj.type === 'gcp') {
if (process.env.CI_PROXY === 'true') {
locationObj.details.https = false;
locationObj.details.proxy = proxyAddress;
}
const connectionAgent = locationObj.details.https ?
new https.Agent({ keepAlive: true }) :
new http.Agent({ keepAlive: true });
const protocol = locationObj.details.https ? 'https' : 'http';
const httpOptions = locationObj.details.proxy ?
{ proxy: locationObj.details.proxy, agent: connectionAgent,
timeout: 0 }
: { agent: connectionAgent, timeout: 0 };
const sslEnabled = locationObj.details.https === true;
// TODO: HTTP requests to AWS are not supported with V4 auth for
// non-file streams which are used by Backbeat. This option will be
// removed once CA certs, proxy setup feature is implemented.
const signatureVersion = !sslEnabled ? 'v2' : 'v4';
const endpoint = locationObj.type === 'gcp' ?
locationObj.details.gcpEndpoint :
locationObj.details.awsEndpoint;
const s3Params = {
endpoint: `${protocol}://${endpoint}`,
debug: false,
// Not implemented yet for streams in node sdk,
// and has no negative impact if stream, so let's
// leave it in for future use
computeChecksums: true,
httpOptions,
// needed for encryption
signatureVersion,
sslEnabled,
maxRetries: 0,
};
// users can either include the desired profile name from their
// ~/.aws/credentials file or include the accessKeyId and
// secretAccessKey directly in the locationConfig
if (locationObj.details.credentialsProfile) {
s3Params.credentials = new AWS.SharedIniFileCredentials({
profile: locationObj.details.credentialsProfile });
} else {
s3Params.accessKeyId =
locationObj.details.credentials.accessKey;
s3Params.secretAccessKey =
locationObj.details.credentials.secretKey;
}
const clientConfig = {
s3Params,
bucketName: locationObj.details.bucketName,
bucketMatch: locationObj.details.bucketMatch,
serverSideEncryption: locationObj.details.serverSideEncryption,
dataStoreName: location,
};
if (locationObj.type === 'gcp') {
clientConfig.overflowBucket =
locationObj.details.overflowBucketName;
clientConfig.mpuBucket = locationObj.details.mpuBucketName;
clientConfig.authParams = config.getGcpServiceParams(location);
}
clients[location] = locationObj.type === 'gcp' ?
new GcpClient(clientConfig) : new AwsClient(clientConfig);
}
if (locationObj.type === 'azure') {
if (process.env.CI_PROXY === 'true') {
locationObj.details.proxy = proxyAddress;
}
const azureStorageEndpoint = config.getAzureEndpoint(location);
const azureStorageCredentials =
config.getAzureStorageCredentials(location);
clients[location] = new AzureClient({
azureStorageEndpoint,
azureStorageCredentials,
azureContainerName: locationObj.details.azureContainerName,
bucketMatch: locationObj.details.bucketMatch,
dataStoreName: location,
proxy: locationObj.details.proxy,
});
clients[location].clientType = 'azure';
}
});
return clients;
}
module.exports = parseLC;

View File

@ -0,0 +1,289 @@
const { errors, s3middleware } = require('arsenal');
const { parseTagFromQuery } = s3middleware.tagging;
const createLogger = require('./multipleBackendLogger');
const async = require('async');
const { config } = require('../Config');
const parseLC = require('./locationConstraintParser');
const { checkExternalBackend } = require('./external/utils');
const { externalBackendHealthCheckInterval } = require('../../constants');
let clients = parseLC(config);
config.on('location-constraints-update', () => {
clients = parseLC(config);
});
const multipleBackendGateway = {
put: (hashedStream, size, keyContext,
backendInfo, reqUids, callback) => {
const controllingLocationConstraint =
backendInfo.getControllingLocationConstraint();
const client = clients[controllingLocationConstraint];
if (!client) {
const log = createLogger(reqUids);
log.error('no data backend matching controlling locationConstraint',
{ controllingLocationConstraint });
return process.nextTick(() => {
callback(errors.InternalError);
});
}
let writeStream = hashedStream;
if (keyContext.cipherBundle && keyContext.cipherBundle.cipher) {
writeStream = keyContext.cipherBundle.cipher;
hashedStream.pipe(writeStream);
}
if (keyContext.tagging) {
const validationTagRes = parseTagFromQuery(keyContext.tagging);
if (validationTagRes instanceof Error) {
const log = createLogger(reqUids);
log.debug('tag validation failed', {
error: validationTagRes,
method: 'multipleBackendGateway put',
});
return callback(errors.InternalError);
}
}
return client.put(writeStream, size, keyContext,
reqUids, (err, key, dataStoreVersionId) => {
const log = createLogger(reqUids);
log.debug('put to location', { controllingLocationConstraint });
if (err) {
log.error('error from datastore',
{ error: err, dataStoreType: client.clientType });
return callback(errors.ServiceUnavailable);
}
const dataRetrievalInfo = {
key,
dataStoreName: controllingLocationConstraint,
dataStoreType: client.clientType,
dataStoreVersionId,
};
return callback(null, dataRetrievalInfo);
});
},
head: (objectGetInfoArr, reqUids, callback) => {
if (!objectGetInfoArr || !Array.isArray(objectGetInfoArr)
|| !objectGetInfoArr[0] || !objectGetInfoArr[0].dataStoreName) {
// no-op if no stored data store name
return process.nextTick(callback);
}
const objectGetInfo = objectGetInfoArr[0];
const client = clients[objectGetInfo.dataStoreName];
if (client.head === undefined) {
// no-op if unsupported client method
return process.nextTick(callback);
}
return client.head(objectGetInfo, reqUids, callback);
},
get: (objectGetInfo, range, reqUids, callback) => {
let key;
let client;
// for backwards compatibility
if (typeof objectGetInfo === 'string') {
key = objectGetInfo;
client = clients.sproxyd;
} else {
key = objectGetInfo.key;
client = clients[objectGetInfo.dataStoreName];
}
if (client.clientType === 'scality') {
return client.get(key, range, reqUids, callback);
}
return client.get(objectGetInfo, range, reqUids, callback);
},
delete: (objectGetInfo, reqUids, callback) => {
let key;
let client;
// for backwards compatibility
if (typeof objectGetInfo === 'string') {
key = objectGetInfo;
client = clients.sproxyd;
} else {
key = objectGetInfo.key;
client = clients[objectGetInfo.dataStoreName];
}
if (client.clientType === 'scality') {
return client.delete(key, reqUids, callback);
}
return client.delete(objectGetInfo, reqUids, callback);
},
healthcheck: (flightCheckOnStartUp, log, callback) => {
const multBackendResp = {};
const awsArray = [];
const azureArray = [];
async.each(Object.keys(clients), (location, cb) => {
const client = clients[location];
if (client.clientType === 'scality') {
return client.healthcheck(log, (err, res) => {
if (err) {
multBackendResp[location] = { error: err };
} else {
multBackendResp[location] = { code: res.statusCode,
message: res.statusMessage };
}
return cb();
});
} else if (client.clientType === 'aws_s3') {
awsArray.push(location);
return cb();
} else if (client.clientType === 'azure') {
azureArray.push(location);
return cb();
}
// if backend type isn't 'scality' or 'aws_s3', it will be
// 'mem' or 'file', for which the default response is 200 OK
multBackendResp[location] = { code: 200, message: 'OK' };
return cb();
}, () => {
async.parallel([
next => checkExternalBackend(
clients, awsArray, 'aws_s3', flightCheckOnStartUp,
externalBackendHealthCheckInterval, next),
next => checkExternalBackend(
clients, azureArray, 'azure', flightCheckOnStartUp,
externalBackendHealthCheckInterval, next),
], (errNull, externalResp) => {
const externalLocResults = [];
externalResp.forEach(resp => externalLocResults.push(...resp));
externalLocResults.forEach(locationResult =>
Object.assign(multBackendResp, locationResult));
callback(null, multBackendResp);
});
});
},
createMPU: (key, metaHeaders, bucketName, websiteRedirectHeader,
location, contentType, cacheControl, contentDisposition,
contentEncoding, log, cb) => {
const client = clients[location];
if (client.clientType === 'aws_s3') {
return client.createMPU(key, metaHeaders, bucketName,
websiteRedirectHeader, contentType, cacheControl,
contentDisposition, contentEncoding, log, cb);
}
return cb();
},
uploadPart: (request, streamingV4Params, stream, size, location, key,
uploadId, partNumber, bucketName, log, cb) => {
const client = clients[location];
if (client.uploadPart) {
return client.uploadPart(request, streamingV4Params, stream, size,
key, uploadId, partNumber, bucketName, log, cb);
}
return cb();
},
listParts: (key, uploadId, location, bucketName, partNumberMarker, maxParts,
log, cb) => {
const client = clients[location];
if (client.listParts) {
return client.listParts(key, uploadId, bucketName, partNumberMarker,
maxParts, log, cb);
}
return cb();
},
completeMPU: (key, uploadId, location, jsonList, mdInfo, bucketName,
userMetadata, contentSettings, log, cb) => {
const client = clients[location];
if (client.completeMPU) {
const args = [jsonList, mdInfo, key, uploadId, bucketName];
if (client.clientType === 'azure') {
args.push(userMetadata, contentSettings);
}
return client.completeMPU(...args, log, (err, completeObjData) => {
if (err) {
return cb(err);
}
// eslint-disable-next-line no-param-reassign
completeObjData.dataStoreType = client.clientType;
return cb(null, completeObjData);
});
}
return cb();
},
abortMPU: (key, uploadId, location, bucketName, log, cb) => {
const client = clients[location];
if (client.clientType === 'azure') {
const skipDataDelete = true;
return cb(null, skipDataDelete);
}
if (client.abortMPU) {
return client.abortMPU(key, uploadId, bucketName, log, err => {
if (err) {
return cb(err);
}
return cb();
});
}
return cb();
},
objectTagging: (method, key, bucket, objectMD, log, cb) => {
// if legacy, objectMD will not contain dataStoreName, so just return
const client = clients[objectMD.dataStoreName];
if (client && client[`object${method}Tagging`]) {
return client[`object${method}Tagging`](key, bucket, objectMD, log,
cb);
}
return cb();
},
// NOTE: using copyObject only if copying object from one external
// backend to the same external backend
copyObject: (request, destLocationConstraintName, externalSourceKey,
sourceLocationConstraintName, storeMetadataParams, log, cb) => {
const client = clients[destLocationConstraintName];
if (client.copyObject) {
return client.copyObject(request, destLocationConstraintName,
externalSourceKey, sourceLocationConstraintName,
storeMetadataParams, log, (err, key, dataStoreVersionId) => {
const dataRetrievalInfo = {
key,
dataStoreName: destLocationConstraintName,
dataStoreType: client.clientType,
dataStoreVersionId,
};
cb(err, dataRetrievalInfo);
});
}
return cb(errors.NotImplemented
.customizeDescription('Can not copy object from ' +
`${client.clientType} to ${client.clientType}`));
},
uploadPartCopy: (request, location, awsSourceKey,
sourceLocationConstraintName, log, cb) => {
const client = clients[location];
if (client.uploadPartCopy) {
return client.uploadPartCopy(request, awsSourceKey,
sourceLocationConstraintName,
log, cb);
}
return cb(errors.NotImplemented.customizeDescription(
'Can not copy object from ' +
`${client.clientType} to ${client.clientType}`));
},
protectAzureBlocks: (bucketName, objectKey, location, log, cb) => {
const client = clients[location];
if (client.protectAzureBlocks) {
return client.protectAzureBlocks(bucketName, objectKey, location,
log, cb);
}
return cb();
},
};
module.exports = multipleBackendGateway;

View File

@ -0,0 +1,11 @@
const werelogs = require('werelogs');
const logger = new werelogs.Logger('MultipleBackendGateway');
function createLogger(reqUids) {
return reqUids ?
logger.newRequestLoggerFromSerializedUids(reqUids) :
logger.newRequestLogger();
}
module.exports = createLogger;

792
backends/data/wrapper.js Normal file
View File

@ -0,0 +1,792 @@
const async = require('async');
const { errors, s3middleware } = require('arsenal');
const PassThrough = require('stream').PassThrough;
const DataFileInterface = require('./file/backend');
const inMemory = require('./in_memory/backend').backend;
const locationConstraintCheck =
require('../api/apiUtils/object/locationConstraintCheck');
const multipleBackendGateway = require('./multipleBackendGateway');
const utils = require('./external/utils');
const { config } = require('../Config');
const MD5Sum = s3middleware.MD5Sum;
const NullStream = s3middleware.NullStream;
const assert = require('assert');
const kms = require('../kms/wrapper');
const externalBackends = require('../../constants').externalBackends;
const constants = require('../../constants');
const { BackendInfo } = require('../api/apiUtils/object/BackendInfo');
const RelayMD5Sum = require('../utilities/RelayMD5Sum');
const skipError = new Error('skip');
let CdmiData;
try {
CdmiData = require('cdmiclient').CdmiData;
} catch (err) {
CdmiData = null;
}
let client;
let implName;
if (config.backends.data === 'mem') {
client = inMemory;
implName = 'mem';
} else if (config.backends.data === 'file') {
client = new DataFileInterface();
implName = 'file';
} else if (config.backends.data === 'multiple') {
client = multipleBackendGateway;
implName = 'multipleBackends';
} else if (config.backends.data === 'cdmi') {
if (!CdmiData) {
throw new Error('Unauthorized backend');
}
client = new CdmiData({
path: config.cdmi.path,
host: config.cdmi.host,
port: config.cdmi.port,
readonly: config.cdmi.readonly,
});
implName = 'cdmi';
}
/**
* _retryDelete - Attempt to delete key again if it failed previously
* @param { string | object } objectGetInfo - either string location of object
* to delete or object containing info of object to delete
* @param {object} log - Werelogs request logger
* @param {number} count - keeps count of number of times function has been run
* @param {function} cb - callback
* @returns undefined and calls callback
*/
const MAX_RETRY = 2;
// This check is done because on a put, complete mpu or copy request to
// Azure/AWS, if the object already exists on that backend, the existing object
// should not be deleted, which is the functionality for all other backends
function _shouldSkipDelete(locations, requestMethod, newObjDataStoreName) {
const skipMethods = { PUT: true, POST: true };
if (!Array.isArray(locations) || !locations[0] ||
!locations[0].dataStoreType) {
return false;
}
const isSkipBackend = externalBackends[locations[0].dataStoreType];
const isMatchingBackends =
locations[0].dataStoreName === newObjDataStoreName;
const isSkipMethod = skipMethods[requestMethod];
return (isSkipBackend && isMatchingBackends && isSkipMethod);
}
function _retryDelete(objectGetInfo, log, count, cb) {
if (count > MAX_RETRY) {
return cb(errors.InternalError);
}
return client.delete(objectGetInfo, log.getSerializedUids(), err => {
if (err) {
log.error('delete error from datastore',
{ error: err, implName, moreRetries: 'yes' });
return _retryDelete(objectGetInfo, log, count + 1, cb);
}
return cb();
});
}
function _put(cipherBundle, value, valueSize,
keyContext, backendInfo, log, cb) {
assert.strictEqual(typeof valueSize, 'number');
log.debug('sending put to datastore', { implName, keyContext,
method: 'put' });
let hashedStream = null;
if (value) {
hashedStream = new MD5Sum();
value.pipe(hashedStream);
}
if (implName === 'multipleBackends') {
// Need to send backendInfo to client.put and
// client.put will provide dataRetrievalInfo so no
// need to construct here
/* eslint-disable no-param-reassign */
keyContext.cipherBundle = cipherBundle;
return client.put(hashedStream,
valueSize, keyContext, backendInfo, log.getSerializedUids(),
(err, dataRetrievalInfo) => {
if (err) {
log.error('put error from datastore',
{ error: err, implName });
return cb(errors.ServiceUnavailable);
}
return cb(null, dataRetrievalInfo, hashedStream);
});
}
/* eslint-enable no-param-reassign */
let writeStream = hashedStream;
if (cipherBundle && cipherBundle.cipher) {
writeStream = cipherBundle.cipher;
hashedStream.pipe(writeStream);
}
return client.put(writeStream, valueSize, keyContext,
log.getSerializedUids(), (err, key) => {
if (err) {
log.error('put error from datastore',
{ error: err, implName });
return cb(errors.InternalError);
}
const dataRetrievalInfo = {
key,
dataStoreName: implName,
};
return cb(null, dataRetrievalInfo, hashedStream);
});
}
const data = {
put: (cipherBundle, value, valueSize, keyContext, backendInfo, log, cb) => {
_put(cipherBundle, value, valueSize, keyContext, backendInfo, log,
(err, dataRetrievalInfo, hashedStream) => {
if (err) {
return cb(err);
}
if (hashedStream) {
if (hashedStream.completedHash) {
return cb(null, dataRetrievalInfo, hashedStream);
}
hashedStream.on('hashed', () => {
hashedStream.removeAllListeners('hashed');
return cb(null, dataRetrievalInfo, hashedStream);
});
return undefined;
}
return cb(null, dataRetrievalInfo);
});
},
head: (objectGetInfo, log, cb) => {
if (implName !== 'multipleBackends') {
// no-op if not multipleBackend implementation;
// head is used during get just to check external backend data state
return process.nextTick(cb);
}
return client.head(objectGetInfo, log.getSerializedUids(), cb);
},
get: (objectGetInfo, response, log, cb) => {
const isMdModelVersion2 = typeof(objectGetInfo) === 'string';
const isRequiredStringKey = constants.clientsRequireStringKey[implName];
const key = isMdModelVersion2 ? objectGetInfo : objectGetInfo.key;
const clientGetInfo = isRequiredStringKey ? key : objectGetInfo;
const range = objectGetInfo.range;
// If the key is explicitly set to null, the part to
// be read doesn't really exist and is only made of zeroes.
// This functionality is used by Scality-NFSD.
// Otherwise, the key is always defined
assert(key === null || key !== undefined);
if (key === null) {
cb(null, new NullStream(objectGetInfo.size, range));
return;
}
log.debug('sending get to datastore', { implName,
key, range, method: 'get' });
// We need to use response as a writable stream for AZURE GET
if (!isMdModelVersion2 && !isRequiredStringKey && response) {
clientGetInfo.response = response;
}
client.get(clientGetInfo, range, log.getSerializedUids(),
(err, stream) => {
if (err) {
log.error('get error from datastore',
{ error: err, implName });
return cb(errors.ServiceUnavailable);
}
if (objectGetInfo.cipheredDataKey) {
const serverSideEncryption = {
cryptoScheme: objectGetInfo.cryptoScheme,
masterKeyId: objectGetInfo.masterKeyId,
cipheredDataKey: Buffer.from(
objectGetInfo.cipheredDataKey, 'base64'),
};
const offset = objectGetInfo.range ?
objectGetInfo.range[0] : 0;
return kms.createDecipherBundle(
serverSideEncryption, offset, log,
(err, decipherBundle) => {
if (err) {
log.error('cannot get decipher bundle ' +
'from kms', {
method: 'data.wrapper.data.get',
});
return cb(err);
}
stream.pipe(decipherBundle.decipher);
return cb(null, decipherBundle.decipher);
});
}
return cb(null, stream);
});
},
delete: (objectGetInfo, log, cb) => {
const callback = cb || log.end;
const isMdModelVersion2 = typeof(objectGetInfo) === 'string';
const isRequiredStringKey = constants.clientsRequireStringKey[implName];
const key = isMdModelVersion2 ? objectGetInfo : objectGetInfo.key;
const clientGetInfo = isRequiredStringKey ? key : objectGetInfo;
log.trace('sending delete to datastore', {
implName, key, method: 'delete' });
// If the key is explicitly set to null, the part to
// be deleted doesn't really exist.
// This functionality is used by Scality-NFSD.
// Otherwise, the key is always defined
assert(key === null || key !== undefined);
if (key === null) {
callback(null);
return;
}
_retryDelete(clientGetInfo, log, 0, err => {
if (err) {
log.error('delete error from datastore',
{ error: err, key: objectGetInfo.key, moreRetries: 'no' });
}
return callback(err);
});
},
// It would be preferable to have an sproxyd batch delete route to
// replace this
batchDelete: (locations, requestMethod, newObjDataStoreName, log) => {
// TODO: The method of persistence of sproxy delete key will
// be finalized; refer Issue #312 for the discussion. In the
// meantime, we at least log the location of the data we are
// about to delete before attempting its deletion.
if (_shouldSkipDelete(locations, requestMethod, newObjDataStoreName)) {
return;
}
log.trace('initiating batch delete', {
keys: locations,
implName,
method: 'batchDelete',
});
async.eachLimit(locations, 5, (loc, next) => {
process.nextTick(() => data.delete(loc, log, next));
},
err => {
if (err) {
log.error('batch delete failed', { error: err });
} else {
log.trace('batch delete successfully completed');
}
log.end();
});
},
switch: newClient => {
client = newClient;
return client;
},
checkHealth: (log, cb, flightCheckOnStartUp) => {
if (!client.healthcheck) {
const defResp = {};
defResp[implName] = { code: 200, message: 'OK' };
return cb(null, defResp);
}
return client.healthcheck(flightCheckOnStartUp, log, (err, result) => {
let respBody = {};
if (err) {
log.error(`error from ${implName}`, { error: err });
respBody[implName] = {
error: err,
};
// error returned as null so async parallel doesn't return
// before all backends are checked
return cb(null, respBody);
}
if (implName === 'multipleBackends') {
respBody = result;
return cb(null, respBody);
}
respBody[implName] = {
code: result.statusCode,
message: result.statusMessage,
};
return cb(null, respBody);
});
},
getDiskUsage: (log, cb) => {
if (!client.getDiskUsage) {
log.debug('returning empty disk usage as fallback', { implName });
return cb(null, {});
}
return client.getDiskUsage(log.getSerializedUids(), cb);
},
/**
* _putForCopy - put used for copying object
* @param {object} cipherBundle - cipher bundle that encrypt the data
* @param {object} stream - stream containing the data
* @param {object} part - element of dataLocator array
* @param {object} dataStoreContext - information of the
* destination object
* dataStoreContext.bucketName: destination bucket name,
* dataStoreContext.owner: owner,
* dataStoreContext.namespace: request namespace,
* dataStoreContext.objectKey: destination object key name,
* @param {BackendInfo} destBackendInfo - Instance of BackendInfo:
* Represents the info necessary to evaluate which data backend to use
* on a data put call.
* @param {object} log - Werelogs request logger
* @param {function} cb - callback
* @returns {function} cb - callback
*/
_putForCopy: (cipherBundle, stream, part, dataStoreContext,
destBackendInfo, log, cb) => data.put(cipherBundle, stream,
part.size, dataStoreContext,
destBackendInfo, log,
(error, partRetrievalInfo) => {
if (error) {
return cb(error);
}
const partResult = {
key: partRetrievalInfo.key,
dataStoreName: partRetrievalInfo
.dataStoreName,
dataStoreType: partRetrievalInfo
.dataStoreType,
start: part.start,
size: part.size,
};
if (cipherBundle) {
partResult.cryptoScheme = cipherBundle.cryptoScheme;
partResult.cipheredDataKey = cipherBundle.cipheredDataKey;
}
if (part.dataStoreETag) {
partResult.dataStoreETag = part.dataStoreETag;
}
if (partRetrievalInfo.dataStoreVersionId) {
partResult.dataStoreVersionId =
partRetrievalInfo.dataStoreVersionId;
}
return cb(null, partResult);
}),
/**
* _dataCopyPut - put used for copying object with and without
* encryption
* @param {string} serverSideEncryption - Server side encryption
* @param {object} stream - stream containing the data
* @param {object} part - element of dataLocator array
* @param {object} dataStoreContext - information of the
* destination object
* dataStoreContext.bucketName: destination bucket name,
* dataStoreContext.owner: owner,
* dataStoreContext.namespace: request namespace,
* dataStoreContext.objectKey: destination object key name,
* @param {BackendInfo} destBackendInfo - Instance of BackendInfo:
* Represents the info necessary to evaluate which data backend to use
* on a data put call.
* @param {object} log - Werelogs request logger
* @param {function} cb - callback
* @returns {function} cb - callback
*/
_dataCopyPut: (serverSideEncryption, stream, part, dataStoreContext,
destBackendInfo, log, cb) => {
if (serverSideEncryption) {
return kms.createCipherBundle(
serverSideEncryption,
log, (err, cipherBundle) => {
if (err) {
log.debug('error getting cipherBundle');
return cb(errors.InternalError);
}
return data._putForCopy(cipherBundle, stream, part,
dataStoreContext, destBackendInfo, log, cb);
});
}
// Copied object is not encrypted so just put it
// without a cipherBundle
return data._putForCopy(null, stream, part, dataStoreContext,
destBackendInfo, log, cb);
},
/**
* copyObject - copy object
* @param {object} request - request object
* @param {string} sourceLocationConstraintName -
* source locationContraint name (awsbackend, azurebackend, ...)
* @param {object} storeMetadataParams - metadata information of the
* source object
* @param {array} dataLocator - source object metadata location(s)
* NOTE: for Azure and AWS data backend this array only has one item
* @param {object} dataStoreContext - information of the
* destination object
* dataStoreContext.bucketName: destination bucket name,
* dataStoreContext.owner: owner,
* dataStoreContext.namespace: request namespace,
* dataStoreContext.objectKey: destination object key name,
* @param {BackendInfo} destBackendInfo - Instance of BackendInfo:
* Represents the info necessary to evaluate which data backend to use
* on a data put call.
* @param {object} sourceBucketMD - metadata of the source bucket
* @param {object} destBucketMD - metadata of the destination bucket
* @param {object} log - Werelogs request logger
* @param {function} cb - callback
* @returns {function} cb - callback
*/
copyObject: (request,
sourceLocationConstraintName, storeMetadataParams, dataLocator,
dataStoreContext, destBackendInfo, sourceBucketMD, destBucketMD, log,
cb) => {
const serverSideEncryption = destBucketMD.getServerSideEncryption();
if (config.backends.data === 'multiple' &&
utils.externalBackendCopy(sourceLocationConstraintName,
storeMetadataParams.dataStoreName, sourceBucketMD, destBucketMD)) {
const destLocationConstraintName =
storeMetadataParams.dataStoreName;
const objectGetInfo = dataLocator[0];
const externalSourceKey = objectGetInfo.key;
return client.copyObject(request, destLocationConstraintName,
externalSourceKey, sourceLocationConstraintName,
storeMetadataParams, log, (error, objectRetrievalInfo) => {
if (error) {
return cb(error);
}
const putResult = {
key: objectRetrievalInfo.key,
dataStoreName: objectRetrievalInfo.
dataStoreName,
dataStoreType: objectRetrievalInfo.
dataStoreType,
dataStoreVersionId:
objectRetrievalInfo.dataStoreVersionId,
size: storeMetadataParams.size,
dataStoreETag: objectGetInfo.dataStoreETag,
start: objectGetInfo.start,
};
const putResultArr = [putResult];
return cb(null, putResultArr);
});
}
// dataLocator is an array. need to get and put all parts
// For now, copy 1 part at a time. Could increase the second
// argument here to increase the number of parts
// copied at once.
return async.mapLimit(dataLocator, 1,
// eslint-disable-next-line prefer-arrow-callback
function copyPart(part, copyCb) {
if (part.dataStoreType === 'azure') {
const passThrough = new PassThrough();
return async.parallel([
parallelCb => data.get(part, passThrough, log, err =>
parallelCb(err)),
parallelCb => data._dataCopyPut(serverSideEncryption,
passThrough,
part, dataStoreContext, destBackendInfo, log,
parallelCb),
], (err, res) => {
if (err) {
return copyCb(err);
}
return copyCb(null, res[1]);
});
}
return data.get(part, null, log, (err, stream) => {
if (err) {
return copyCb(err);
}
return data._dataCopyPut(serverSideEncryption, stream,
part, dataStoreContext, destBackendInfo, log, copyCb);
});
}, (err, results) => {
if (err) {
log.debug('error transferring data from source',
{ error: err });
return cb(err);
}
return cb(null, results);
});
},
_dataCopyPutPart: (request,
serverSideEncryption, stream, part,
dataStoreContext, destBackendInfo, locations, log, cb) => {
const numberPartSize =
Number.parseInt(part.size, 10);
const partNumber = Number.parseInt(request.query.partNumber, 10);
const uploadId = request.query.uploadId;
const destObjectKey = request.objectKey;
const destBucketName = request.bucketName;
const destLocationConstraintName = destBackendInfo
.getControllingLocationConstraint();
if (externalBackends[config
.locationConstraints[destLocationConstraintName]
.type]) {
return multipleBackendGateway.uploadPart(null, null,
stream, numberPartSize,
destLocationConstraintName, destObjectKey, uploadId,
partNumber, destBucketName, log,
(err, partInfo) => {
if (err) {
log.error('error putting ' +
'part to AWS', {
error: err,
method:
'objectPutCopyPart::' +
'multipleBackendGateway.' +
'uploadPart',
});
return cb(errors.ServiceUnavailable);
}
// skip to end of waterfall
// because don't need to store
// part metadata
if (partInfo &&
partInfo.dataStoreType === 'aws_s3') {
// if data backend handles MPU, skip to end
// of waterfall
const partResult = {
dataStoreETag: partInfo.dataStoreETag,
};
locations.push(partResult);
return cb(skipError, partInfo.dataStoreETag);
} else if (
partInfo &&
partInfo.dataStoreType === 'azure') {
const partResult = {
key: partInfo.key,
dataStoreName: partInfo.dataStoreName,
dataStoreETag: partInfo.dataStoreETag,
size: numberPartSize,
numberSubParts:
partInfo.numberSubParts,
partNumber: partInfo.partNumber,
};
locations.push(partResult);
return cb();
}
return cb(skipError);
});
}
if (serverSideEncryption) {
return kms.createCipherBundle(
serverSideEncryption,
log, (err, cipherBundle) => {
if (err) {
log.debug('error getting cipherBundle',
{ error: err });
return cb(errors.InternalError);
}
return data.put(cipherBundle, stream,
numberPartSize, dataStoreContext,
destBackendInfo, log,
(error, partRetrievalInfo,
hashedStream) => {
if (error) {
log.debug('error putting ' +
'encrypted part', { error });
return cb(error);
}
const partResult = {
key: partRetrievalInfo.key,
dataStoreName: partRetrievalInfo
.dataStoreName,
dataStoreETag: hashedStream
.completedHash,
// Do not include part start
// here since will change in
// final MPU object
size: numberPartSize,
sseCryptoScheme: cipherBundle
.cryptoScheme,
sseCipheredDataKey: cipherBundle
.cipheredDataKey,
sseAlgorithm: cipherBundle
.algorithm,
sseMasterKeyId: cipherBundle
.masterKeyId,
};
locations.push(partResult);
return cb();
});
});
}
// Copied object is not encrypted so just put it
// without a cipherBundle
return data.put(null, stream, numberPartSize,
dataStoreContext, destBackendInfo,
log, (error, partRetrievalInfo, hashedStream) => {
if (error) {
log.debug('error putting object part',
{ error });
return cb(error);
}
const partResult = {
key: partRetrievalInfo.key,
dataStoreName: partRetrievalInfo.dataStoreName,
dataStoreETag: hashedStream.completedHash,
size: numberPartSize,
};
locations.push(partResult);
return cb();
});
},
/**
* uploadPartCopy - put copy part
* @param {object} request - request object
* @param {object} log - Werelogs request logger
* @param {object} destBucketMD - destination bucket metadata
* @param {string} sourceLocationConstraintName -
* source locationContraint name (awsbackend, azurebackend, ...)
* @param {string} destLocationConstraintName -
* location of the destination MPU object (awsbackend, azurebackend, ...)
* @param {array} dataLocator - source object metadata location(s)
* NOTE: for Azure and AWS data backend this array
* @param {object} dataStoreContext - information of the
* destination object
* dataStoreContext.bucketName: destination bucket name,
* dataStoreContext.owner: owner,
* dataStoreContext.namespace: request namespace,
* dataStoreContext.objectKey: destination object key name,
* dataStoreContext.uploadId: uploadId
* dataStoreContext.partNumber: request.query.partNumber
* @param {function} callback - callback
* @returns {function} cb - callback
*/
uploadPartCopy: (request, log, destBucketMD, sourceLocationConstraintName,
destLocationConstraintName, dataLocator, dataStoreContext,
callback) => {
const serverSideEncryption = destBucketMD.getServerSideEncryption();
const lastModified = new Date().toJSON();
// skip if 0 byte object
if (dataLocator.length === 0) {
return process.nextTick(() => {
callback(null, constants.emptyFileMd5,
lastModified, serverSideEncryption, []);
});
}
// if destination mpu was initiated in legacy version
if (destLocationConstraintName === undefined) {
const backendInfoObj = locationConstraintCheck(request,
null, destBucketMD, log);
if (backendInfoObj.err) {
return process.nextTick(() => {
callback(backendInfoObj.err);
});
}
// eslint-disable-next-line no-param-reassign
destLocationConstraintName = backendInfoObj.controllingLC;
}
const locationTypeMatchAWS =
config.backends.data === 'multiple' &&
config.getLocationConstraintType(sourceLocationConstraintName) ===
config.getLocationConstraintType(destLocationConstraintName) &&
config.getLocationConstraintType(sourceLocationConstraintName) ===
'aws_s3';
// NOTE: using multipleBackendGateway.uploadPartCopy only if copying
// from AWS to AWS
if (locationTypeMatchAWS && dataLocator.length === 1) {
const awsSourceKey = dataLocator[0].key;
return multipleBackendGateway.uploadPartCopy(request,
destLocationConstraintName, awsSourceKey,
sourceLocationConstraintName, log, (error, eTag) => {
if (error) {
return callback(error);
}
return callback(skipError, eTag,
lastModified, serverSideEncryption);
});
}
const backendInfo = new BackendInfo(destLocationConstraintName);
// totalHash will be sent through the RelayMD5Sum transform streams
// to collect the md5 from multiple streams
let totalHash;
const locations = [];
// dataLocator is an array. need to get and put all parts
// in order so can get the ETag of full object
return async.forEachOfSeries(dataLocator,
// eslint-disable-next-line prefer-arrow-callback
function copyPart(part, index, cb) {
if (part.dataStoreType === 'azure') {
const passThrough = new PassThrough();
return async.parallel([
next => data.get(part, passThrough, log, err => {
if (err) {
log.error('error getting data part ' +
'from Azure', {
error: err,
method:
'objectPutCopyPart::' +
'multipleBackendGateway.' +
'copyPart',
});
return next(err);
}
return next();
}),
next => data._dataCopyPutPart(request,
serverSideEncryption, passThrough, part,
dataStoreContext, backendInfo, locations, log, next),
], err => {
if (err) {
return cb(err);
}
return cb();
});
}
return data.get(part, null, log, (err, stream) => {
if (err) {
log.debug('error getting object part',
{ error: err });
return cb(err);
}
const hashedStream =
new RelayMD5Sum(totalHash, updatedHash => {
totalHash = updatedHash;
});
stream.pipe(hashedStream);
// destLocationConstraintName is location of the
// destination MPU object
return data._dataCopyPutPart(request,
serverSideEncryption, hashedStream, part,
dataStoreContext, backendInfo, locations, log, cb);
});
}, err => {
// Digest the final combination of all of the part streams
if (err && err !== skipError) {
log.debug('error transferring data from source',
{ error: err, method: 'goGetData' });
return callback(err);
}
if (totalHash) {
totalHash = totalHash.digest('hex');
} else {
totalHash = locations[0].dataStoreETag;
}
if (err && err === skipError) {
return callback(skipError, totalHash,
lastModified, serverSideEncryption);
}
return callback(null, totalHash,
lastModified, serverSideEncryption, locations);
});
},
};
module.exports = data;

View File

@ -0,0 +1,82 @@
# BucketInfo Model Version History
## Model Version 0/1
### Properties
``` javascript
this._acl = aclInstance;
this._name = name;
this._owner = owner;
this._ownerDisplayName = ownerDisplayName;
this._creationDate = creationDate;
```
### Usage
No explicit references in the code since mdBucketModelVersion
property not added until Model Version 2
## Model Version 2
### Properties Added
``` javascript
this._mdBucketModelVersion = mdBucketModelVersion || 0
this._transient = transient || false;
this._deleted = deleted || false;
```
### Usage
Used to determine which splitter to use ( < 2 means old splitter)
## Model version 3
### Properties Added
```
this._serverSideEncryption = serverSideEncryption || null;
```
### Usage
Used to store the server bucket encryption info
## Model version 4
### Properties Added
```javascript
this._locationConstraint = LocationConstraint || null;
```
### Usage
Used to store the location constraint of the bucket
## Model version 5
### Properties Added
```javascript
this._websiteConfiguration = websiteConfiguration || null;
this._cors = cors || null;
```
### Usage
Used to store the bucket website configuration info
and to store CORS rules to apply to cross-domain requests
## Model version 6
### Properties Added
```javascript
this._lifecycleConfiguration = lifecycleConfiguration || null;
```
### Usage
Used to store the bucket lifecycle configuration info

View File

@ -1,5 +1,6 @@
const { errors } = require('arsenal');
const getReplicationInfo = require('../api/apiUtils/object/getReplicationInfo');
const aclUtils = require('../utilities/aclUtils');
const constants = require('../../constants');
const metadata = require('../metadata/wrapper');
@ -12,40 +13,18 @@ const acl = {
metadata.updateBucket(bucket.getName(), bucket, log, cb);
},
/**
* returns true if the specified ACL grant is unchanged
* @param {string} grant name of the grant
* @param {object} oldAcl old acl config
* @param {object} newAcl new acl config
* @returns {bool} is the grant the same
*/
_aclGrantDidNotChange(grant, oldAcl, newAcl) {
if (grant === 'Canned') {
return oldAcl.Canned === newAcl.Canned;
}
/**
* An ACL grant is in form of an array of strings
* An ACL grant is considered unchanged when both the old and new one
* contain the same number of elements, and all elements from one
* grant are incuded in the other grant
*/
return oldAcl[grant].length === newAcl[grant].length
&& oldAcl[grant].every(value => newAcl[grant].includes(value));
},
addObjectACL(bucket, objectKey, objectMD, addACLParams, params, log, cb) {
log.trace('updating object acl in metadata');
const isAclUnchanged = Object.keys(objectMD.acl).length === Object.keys(addACLParams).length
&& Object.keys(objectMD.acl).every(grant => this._aclGrantDidNotChange(grant, objectMD.acl, addACLParams));
if (!isAclUnchanged) {
/* eslint-disable no-param-reassign */
objectMD.acl = addACLParams;
objectMD.originOp = 's3:ObjectAcl:Put';
/* eslint-disable no-param-reassign */
return metadata.putObjectMD(bucket.getName(), objectKey, objectMD, params, log,
cb);
// eslint-disable-next-line no-param-reassign
objectMD.acl = addACLParams;
const replicationInfo = getReplicationInfo(objectKey, bucket, true);
if (replicationInfo) {
// eslint-disable-next-line no-param-reassign
objectMD.replicationInfo = Object.assign({},
objectMD.replicationInfo, replicationInfo);
}
return cb();
metadata.putObjectMD(bucket.getName(), objectKey, objectMD, params, log,
cb);
},
parseAclFromHeaders(params, cb) {
@ -157,4 +136,3 @@ const acl = {
};
module.exports = acl;

View File

@ -0,0 +1,213 @@
const assert = require('assert');
const bucketclient = require('bucketclient');
const BucketInfo = require('arsenal').models.BucketInfo;
const logger = require('../../utilities/logger');
const { config } = require('../../Config');
class BucketClientInterface {
constructor() {
assert(config.bucketd.bootstrap.length > 0,
'bucketd bootstrap list is empty');
const { bootstrap, log } = config.bucketd;
if (config.https) {
const { key, cert, ca } = config.https;
logger.info('bucketclient configuration', {
bootstrap,
log,
https: true,
});
this.client = new bucketclient.RESTClient(bootstrap, log, true,
key, cert, ca);
} else {
logger.info('bucketclient configuration', {
bootstrap,
log,
https: false,
});
this.client = new bucketclient.RESTClient(bootstrap, log);
}
}
createBucket(bucketName, bucketMD, log, cb) {
this.client.createBucket(bucketName, log.getSerializedUids(),
bucketMD.serialize(), cb);
return null;
}
getBucketAttributes(bucketName, log, cb) {
this.client.getBucketAttributes(bucketName, log.getSerializedUids(),
(err, data) => {
if (err) {
return cb(err);
}
return cb(err, BucketInfo.deSerialize(data));
});
return null;
}
getBucketAndObject(bucketName, objName, params, log, cb) {
this.client.getBucketAndObject(bucketName, objName,
log.getSerializedUids(), (err, data) => {
if (err && (!err.NoSuchKey && !err.ObjNotFound)) {
return cb(err);
}
return cb(null, JSON.parse(data));
}, params);
return null;
}
getRaftBuckets(raftId, log, cb) {
return this.client.getRaftBuckets(raftId, log.getSerializedUids(),
(err, data) => {
if (err) {
return cb(err);
}
return cb(null, JSON.parse(data));
});
}
putBucketAttributes(bucketName, bucketMD, log, cb) {
this.client.putBucketAttributes(bucketName, log.getSerializedUids(),
bucketMD.serialize(), cb);
return null;
}
deleteBucket(bucketName, log, cb) {
this.client.deleteBucket(bucketName, log.getSerializedUids(), cb);
return null;
}
putObject(bucketName, objName, objVal, params, log, cb) {
this.client.putObject(bucketName, objName, JSON.stringify(objVal),
log.getSerializedUids(), cb, params);
return null;
}
getObject(bucketName, objName, params, log, cb) {
this.client.getObject(bucketName, objName, log.getSerializedUids(),
(err, data) => {
if (err) {
return cb(err);
}
return cb(err, JSON.parse(data));
}, params);
return null;
}
deleteObject(bucketName, objName, params, log, cb) {
this.client.deleteObject(bucketName, objName, log.getSerializedUids(),
cb, params);
return null;
}
listObject(bucketName, params, log, cb) {
this.client.listObject(bucketName, log.getSerializedUids(), params,
(err, data) => {
if (err) {
return cb(err);
}
return cb(err, JSON.parse(data));
});
return null;
}
listMultipartUploads(bucketName, params, log, cb) {
this.client.listObject(bucketName, log.getSerializedUids(), params,
(err, data) => {
if (err) {
return cb(err);
}
return cb(null, JSON.parse(data));
});
return null;
}
_analyzeHealthFailure(log, callback) {
let doFail = false;
const reason = {
msg: 'Map is available and failure ratio is acceptable',
};
// The healthCheck exposed by Bucketd is a light one, we need
// to inspect all the RaftSession's statuses to make sense of
// it:
return this.client.getAllRafts(undefined, (error, payload) => {
let statuses = null;
try {
statuses = JSON.parse(payload);
} catch (e) {
doFail = true;
reason.msg = 'could not interpret status: invalid payload';
// Can't do anything anymore if we fail here. return.
return callback(doFail, reason);
}
const reducer = (acc, payload) => acc + !payload.connectedToLeader;
reason.ratio = statuses.reduce(reducer, 0) / statuses.length;
/* NOTE FIXME/TODO: acceptableRatio could be configured later on */
reason.acceptableRatio = 0.5;
/* If the RaftSession 0 (map) does not work, fail anyways */
if (!doFail && !statuses[0].connectedToLeader) {
doFail = true;
reason.msg = 'Bucket map unavailable';
}
if (!doFail && reason.ratio > reason.acceptableRatio) {
doFail = true;
reason.msg = 'Ratio of failing Raft Sessions is too high';
}
return callback(doFail, reason);
}, log);
}
/*
* Bucketd offers a behavior that diverges from other sub-components on the
* healthCheck: If any of the pieces making up the bucket storage fail (ie:
* if any Raft Session is down), bucketd returns a 500 for the healthCheck.
*
* As seen in S3C-1412, this may become an issue for S3, whenever the
* system is only partly failing.
*
* This means that S3 needs to analyze the situation, and interpret this
* status depending on the analysis. S3 will then assess the situation as
* critical (and consider it a failure), or not (and consider it a success
* anyways, thus not diverging from the healthCheck behavior of other
* components).
*/
checkHealth(implName, log, cb) {
return this.client.healthcheck(log, (err, result) => {
const respBody = {};
if (err) {
return this._analyzeHealthFailure(log, (failure, reason) => {
const message = reason.msg;
// Remove 'msg' from the reason payload.
// eslint-disable-next-line no-param-reassign
reason.msg = undefined;
respBody[implName] = {
code: 200,
message, // Provide interpreted reason msg
body: reason, // Provide analysis data
};
if (failure) {
// Setting the `error` field is how the healthCheck
// logic interprets it as an error. Don't forget it !
respBody[implName].error = err;
respBody[implName].code = err.code; // original error
}
// error returned as null so async parallel doesn't return
// before all backends are checked
return cb(null, respBody);
}, log);
}
const parseResult = JSON.parse(result);
respBody[implName] = {
code: 200,
message: 'OK',
body: parseResult,
};
return cb(null, respBody);
});
}
}
module.exports = BucketClientInterface;

View File

@ -0,0 +1,379 @@
const cluster = require('cluster');
const arsenal = require('arsenal');
const logger = require('../../utilities/logger');
const BucketInfo = arsenal.models.BucketInfo;
const constants = require('../../../constants');
const { config } = require('../../Config');
const errors = arsenal.errors;
const MetadataFileClient = arsenal.storage.metadata.MetadataFileClient;
const versionSep = arsenal.versioning.VersioningConstants.VersionId.Separator;
const METASTORE = '__metastore';
class BucketFileInterface {
/**
* @constructor
* @param {object} [params] - constructor params
* @param {boolean} [params.noDbOpen=false] - true to skip DB open
* (for unit tests only)
*/
constructor(params) {
this.logger = logger;
const { host, port } = config.metadataClient;
this.mdClient = new MetadataFileClient({ host, port });
if (params && params.noDbOpen) {
return;
}
this.mdDB = this.mdClient.openDB(err => {
if (err) {
throw err;
}
// the metastore sublevel is used to store bucket attributes
this.metastore = this.mdDB.openSub(METASTORE);
if (cluster.isMaster) {
this.setupMetadataServer();
}
});
}
setupMetadataServer() {
/* Since the bucket creation API is expecting the
usersBucket to have attributes, we pre-create the
usersBucket attributes here */
this.mdClient.logger.debug('setting up metadata server');
const usersBucketAttr = new BucketInfo(constants.usersBucket,
'admin', 'admin', new Date().toJSON(),
BucketInfo.currentModelVersion());
this.metastore.put(
constants.usersBucket,
usersBucketAttr.serialize(), {}, err => {
if (err) {
this.logger.fatal('error writing usersBucket ' +
'attributes to metadata',
{ error: err });
throw (errors.InternalError);
}
});
}
/**
* Load DB if exists
* @param {String} bucketName - name of bucket
* @param {Object} log - logger
* @param {function} cb - callback(err, db, attr)
* @return {undefined}
*/
loadDBIfExists(bucketName, log, cb) {
this.getBucketAttributes(bucketName, log, (err, attr) => {
if (err) {
return cb(err);
}
try {
const db = this.mdDB.openSub(bucketName);
return cb(null, db, attr);
} catch (err) {
return cb(errors.InternalError);
}
});
return undefined;
}
createBucket(bucketName, bucketMD, log, cb) {
this.getBucketAttributes(bucketName, log, err => {
if (err && err !== errors.NoSuchBucket) {
return cb(err);
}
if (err === undefined) {
return cb(errors.BucketAlreadyExists);
}
this.putBucketAttributes(bucketName,
bucketMD,
log, cb);
return undefined;
});
}
getBucketAttributes(bucketName, log, cb) {
this.metastore
.withRequestLogger(log)
.get(bucketName, {}, (err, data) => {
if (err) {
if (err.ObjNotFound) {
return cb(errors.NoSuchBucket);
}
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error getting db attributes', logObj);
return cb(errors.InternalError);
}
return cb(null, BucketInfo.deSerialize(data));
});
return undefined;
}
getBucketAndObject(bucketName, objName, params, log, cb) {
this.loadDBIfExists(bucketName, log, (err, db, bucketAttr) => {
if (err) {
return cb(err);
}
db.withRequestLogger(log)
.get(objName, params, (err, objAttr) => {
if (err) {
if (err.ObjNotFound) {
return cb(null, {
bucket: bucketAttr.serialize(),
});
}
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error getting object', logObj);
return cb(errors.InternalError);
}
return cb(null, {
bucket: bucketAttr.serialize(),
obj: objAttr,
});
});
return undefined;
});
return undefined;
}
putBucketAttributes(bucketName, bucketMD, log, cb) {
this.metastore
.withRequestLogger(log)
.put(bucketName, bucketMD.serialize(), {}, err => {
if (err) {
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error putting db attributes', logObj);
return cb(errors.InternalError);
}
return cb();
});
return undefined;
}
deleteBucket(bucketName, log, cb) {
this.metastore
.withRequestLogger(log)
.del(bucketName, {}, err => {
if (err) {
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error deleting bucket',
logObj);
return cb(errors.InternalError);
}
return cb();
});
return undefined;
}
putObject(bucketName, objName, objVal, params, log, cb) {
this.loadDBIfExists(bucketName, log, (err, db) => {
if (err) {
return cb(err);
}
db.withRequestLogger(log)
.put(objName, JSON.stringify(objVal), params, (err, data) => {
if (err) {
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error putting object', logObj);
return cb(errors.InternalError);
}
return cb(err, data);
});
return undefined;
});
}
getObject(bucketName, objName, params, log, cb) {
this.loadDBIfExists(bucketName, log, (err, db) => {
if (err) {
return cb(err);
}
db.withRequestLogger(log).get(objName, params, (err, data) => {
if (err) {
if (err.ObjNotFound) {
return cb(errors.NoSuchKey);
}
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error getting object', logObj);
return cb(errors.InternalError);
}
return cb(null, JSON.parse(data));
});
return undefined;
});
}
deleteObject(bucketName, objName, params, log, cb) {
this.loadDBIfExists(bucketName, log, (err, db) => {
if (err) {
return cb(err);
}
db.withRequestLogger(log).del(objName, params, err => {
if (err) {
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error deleting object', logObj);
return cb(errors.InternalError);
}
return cb();
});
return undefined;
});
}
/**
* This complex function deals with different extensions of bucket listing:
* Delimiter based search or MPU based search.
* @param {String} bucketName - The name of the bucket to list
* @param {Object} params - The params to search
* @param {Object} log - The logger object
* @param {function} cb - Callback when done
* @return {undefined}
*/
internalListObject(bucketName, params, log, cb) {
const extName = params.listingType;
const extension = new arsenal.algorithms.list[extName](params, log);
const requestParams = extension.genMDParams();
this.loadDBIfExists(bucketName, log, (err, db) => {
if (err) {
return cb(err);
}
let cbDone = false;
db.withRequestLogger(log)
.createReadStream(requestParams, (err, stream) => {
if (err) {
return cb(err);
}
stream
.on('data', e => {
if (extension.filter(e) < 0) {
stream.emit('end');
stream.destroy();
}
})
.on('error', err => {
if (!cbDone) {
cbDone = true;
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error listing objects', logObj);
cb(errors.InternalError);
}
})
.on('end', () => {
if (!cbDone) {
cbDone = true;
const data = extension.result();
cb(null, data);
}
});
return undefined;
});
return undefined;
});
}
listObject(bucketName, params, log, cb) {
return this.internalListObject(bucketName, params, log, cb);
}
listMultipartUploads(bucketName, params, log, cb) {
return this.internalListObject(bucketName, params, log, cb);
}
getUUID(log, cb) {
return this.mdDB.getUUID(cb);
}
getDiskUsage(cb) {
return this.mdDB.getDiskUsage(cb);
}
countItems(log, cb) {
const params = {};
const extension = new arsenal.algorithms.list.Basic(params, log);
const requestParams = extension.genMDParams();
const res = {
objects: 0,
versions: 0,
buckets: 0,
};
let cbDone = false;
this.mdDB.rawListKeys(requestParams, (err, stream) => {
if (err) {
return cb(err);
}
stream
.on('data', e => {
if (!e.includes(METASTORE)) {
if (e.includes(constants.usersBucket)) {
res.buckets++;
} else if (e.includes(versionSep)) {
res.versions++;
} else {
res.objects++;
}
}
})
.on('error', err => {
if (!cbDone) {
cbDone = true;
const logObj = {
error: err,
errorMessage: err.message,
errorStack: err.stack,
};
log.error('error listing objects', logObj);
cb(errors.InternalError);
}
})
.on('end', () => {
if (!cbDone) {
cbDone = true;
return cb(null, res);
}
return undefined;
});
return undefined;
});
return undefined;
}
}
module.exports = BucketFileInterface;

View File

@ -0,0 +1,34 @@
const ListResult = require('./ListResult');
class ListMultipartUploadsResult extends ListResult {
constructor() {
super();
this.Uploads = [];
this.NextKeyMarker = undefined;
this.NextUploadIdMarker = undefined;
}
addUpload(uploadInfo) {
this.Uploads.push({
key: decodeURIComponent(uploadInfo.key),
value: {
UploadId: uploadInfo.uploadId,
Initiator: {
ID: uploadInfo.initiatorID,
DisplayName: uploadInfo.initiatorDisplayName,
},
Owner: {
ID: uploadInfo.ownerID,
DisplayName: uploadInfo.ownerDisplayName,
},
StorageClass: uploadInfo.storageClass,
Initiated: uploadInfo.initiated,
},
});
this.MaxKeys += 1;
}
}
module.exports = {
ListMultipartUploadsResult,
};

View File

@ -0,0 +1,27 @@
class ListResult {
constructor() {
this.IsTruncated = false;
this.NextMarker = undefined;
this.CommonPrefixes = [];
/*
Note: this.MaxKeys will get incremented as
keys are added so that when response is returned,
this.MaxKeys will equal total keys in response
(with each CommonPrefix counting as 1 key)
*/
this.MaxKeys = 0;
}
addCommonPrefix(prefix) {
if (!this.hasCommonPrefix(prefix)) {
this.CommonPrefixes.push(prefix);
this.MaxKeys += 1;
}
}
hasCommonPrefix(prefix) {
return (this.CommonPrefixes.indexOf(prefix) !== -1);
}
}
module.exports = ListResult;

View File

@ -0,0 +1,331 @@
const { errors, algorithms, versioning } = require('arsenal');
const getMultipartUploadListing = require('./getMultipartUploadListing');
const { metadata } = require('./metadata');
const { config } = require('../../Config');
const genVID = versioning.VersionID.generateVersionId;
const defaultMaxKeys = 1000;
let uidCounter = 0;
function generateVersionId() {
return genVID(uidCounter++, config.replicationGroupId);
}
function formatVersionKey(key, versionId) {
return `${key}\0${versionId}`;
}
function inc(str) {
return str ? (str.slice(0, str.length - 1) +
String.fromCharCode(str.charCodeAt(str.length - 1) + 1)) : str;
}
const metastore = {
createBucket: (bucketName, bucketMD, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, (err, bucket) => {
// TODO Check whether user already owns the bucket,
// if so return "BucketAlreadyOwnedByYou"
// If not owned by user, return "BucketAlreadyExists"
if (bucket) {
return cb(errors.BucketAlreadyExists);
}
metadata.buckets.set(bucketName, bucketMD);
metadata.keyMaps.set(bucketName, new Map);
return cb();
});
});
},
putBucketAttributes: (bucketName, bucketMD, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, err => {
if (err) {
return cb(err);
}
metadata.buckets.set(bucketName, bucketMD);
return cb();
});
});
},
getBucketAttributes: (bucketName, log, cb) => {
process.nextTick(() => {
if (!metadata.buckets.has(bucketName)) {
return cb(errors.NoSuchBucket);
}
return cb(null, metadata.buckets.get(bucketName));
});
},
deleteBucket: (bucketName, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, err => {
if (err) {
return cb(err);
}
if (metadata.keyMaps.has(bucketName)
&& metadata.keyMaps.get(bucketName).length > 0) {
return cb(errors.BucketNotEmpty);
}
metadata.buckets.delete(bucketName);
metadata.keyMaps.delete(bucketName);
return cb(null);
});
});
},
putObject: (bucketName, objName, objVal, params, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, err => {
if (err) {
return cb(err);
}
/*
valid combinations of versioning options:
- !versioning && !versionId: normal non-versioning put
- versioning && !versionId: create a new version
- versionId: update (PUT/DELETE) an existing version,
and also update master version in case the put
version is newer or same version than master.
if versionId === '' update master version
*/
if (params && params.versionId) {
objVal.versionId = params.versionId; // eslint-disable-line
const mst = metadata.keyMaps.get(bucketName).get(objName);
if (mst && mst.versionId === params.versionId || !mst) {
metadata.keyMaps.get(bucketName).set(objName, objVal);
}
// eslint-disable-next-line
objName = formatVersionKey(objName, params.versionId);
metadata.keyMaps.get(bucketName).set(objName, objVal);
return cb(null, `{"versionId":"${objVal.versionId}"}`);
}
if (params && params.versioning) {
const versionId = generateVersionId();
objVal.versionId = versionId; // eslint-disable-line
metadata.keyMaps.get(bucketName).set(objName, objVal);
// eslint-disable-next-line
objName = formatVersionKey(objName, versionId);
metadata.keyMaps.get(bucketName).set(objName, objVal);
return cb(null, `{"versionId":"${versionId}"}`);
}
if (params && params.versionId === '') {
const versionId = generateVersionId();
objVal.versionId = versionId; // eslint-disable-line
metadata.keyMaps.get(bucketName).set(objName, objVal);
return cb(null, `{"versionId":"${objVal.versionId}"}`);
}
metadata.keyMaps.get(bucketName).set(objName, objVal);
return cb(null);
});
});
},
getBucketAndObject: (bucketName, objName, params, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, (err, bucket) => {
if (err) {
return cb(err, { bucket });
}
if (params && params.versionId) {
// eslint-disable-next-line
objName = formatVersionKey(objName, params.versionId);
}
if (!metadata.keyMaps.has(bucketName)
|| !metadata.keyMaps.get(bucketName).has(objName)) {
return cb(null, { bucket: bucket.serialize() });
}
return cb(null, {
bucket: bucket.serialize(),
obj: JSON.stringify(
metadata.keyMaps.get(bucketName).get(objName)
),
});
});
});
},
getObject: (bucketName, objName, params, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, err => {
if (err) {
return cb(err);
}
if (params && params.versionId) {
// eslint-disable-next-line
objName = formatVersionKey(objName, params.versionId);
}
if (!metadata.keyMaps.has(bucketName)
|| !metadata.keyMaps.get(bucketName).has(objName)) {
return cb(errors.NoSuchKey);
}
return cb(null, metadata.keyMaps.get(bucketName).get(objName));
});
});
},
deleteObject: (bucketName, objName, params, log, cb) => {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, err => {
if (err) {
return cb(err);
}
if (!metadata.keyMaps.get(bucketName).has(objName)) {
return cb(errors.NoSuchKey);
}
if (params && params.versionId) {
const baseKey = inc(formatVersionKey(objName, ''));
const vobjName = formatVersionKey(objName,
params.versionId);
metadata.keyMaps.get(bucketName).delete(vobjName);
const mst = metadata.keyMaps.get(bucketName).get(objName);
if (mst.versionId === params.versionId) {
const keys = [];
metadata.keyMaps.get(bucketName).forEach((val, key) => {
if (key < baseKey && key > vobjName) {
keys.push(key);
}
});
if (keys.length === 0) {
metadata.keyMaps.get(bucketName).delete(objName);
return cb();
}
const key = keys.sort()[0];
const value = metadata.keyMaps.get(bucketName).get(key);
metadata.keyMaps.get(bucketName).set(objName, value);
}
return cb();
}
metadata.keyMaps.get(bucketName).delete(objName);
return cb();
});
});
},
_hasDeleteMarker(key, keyMap) {
const objectMD = keyMap.get(key);
if (objectMD['x-amz-delete-marker'] !== undefined) {
return (objectMD['x-amz-delete-marker'] === true);
}
return false;
},
listObject(bucketName, params, log, cb) {
process.nextTick(() => {
const {
prefix,
marker,
delimiter,
maxKeys,
continuationToken,
startAfter,
} = params;
if (prefix && typeof prefix !== 'string') {
return cb(errors.InvalidArgument);
}
if (marker && typeof marker !== 'string') {
return cb(errors.InvalidArgument);
}
if (delimiter && typeof delimiter !== 'string') {
return cb(errors.InvalidArgument);
}
if (maxKeys && typeof maxKeys !== 'number') {
return cb(errors.InvalidArgument);
}
if (continuationToken && typeof continuationToken !== 'string') {
return cb(errors.InvalidArgument);
}
if (startAfter && typeof startAfter !== 'string') {
return cb(errors.InvalidArgument);
}
// If paramMaxKeys is undefined, the default parameter will set it.
// However, if it is null, the default parameter will not set it.
let numKeys = maxKeys;
if (numKeys === null) {
numKeys = defaultMaxKeys;
}
if (!metadata.keyMaps.has(bucketName)) {
return cb(errors.NoSuchBucket);
}
// If marker specified, edit the keys array so it
// only contains keys that occur alphabetically after the marker
const listingType = params.listingType;
const extension = new algorithms.list[listingType](params, log);
const listingParams = extension.genMDParams();
const keys = [];
metadata.keyMaps.get(bucketName).forEach((val, key) => {
if (listingParams.gt && listingParams.gt >= key) {
return null;
}
if (listingParams.gte && listingParams.gte > key) {
return null;
}
if (listingParams.lt && key >= listingParams.lt) {
return null;
}
if (listingParams.lte && key > listingParams.lte) {
return null;
}
return keys.push(key);
});
keys.sort();
// Iterate through keys array and filter keys containing
// delimiter into response.CommonPrefixes and filter remaining
// keys into response.Contents
for (let i = 0; i < keys.length; ++i) {
const currentKey = keys[i];
// Do not list object with delete markers
if (this._hasDeleteMarker(currentKey,
metadata.keyMaps.get(bucketName))) {
continue;
}
const objMD = metadata.keyMaps.get(bucketName).get(currentKey);
const value = JSON.stringify(objMD);
const obj = {
key: currentKey,
value,
};
// calling Ext.filter(obj) adds the obj to the Ext result if
// not filtered.
// Also, Ext.filter returns false when hit max keys.
// What a nifty function!
if (extension.filter(obj) < 0) {
break;
}
}
return cb(null, extension.result());
});
},
listMultipartUploads(bucketName, listingParams, log, cb) {
process.nextTick(() => {
metastore.getBucketAttributes(bucketName, log, (err, bucket) => {
if (bucket === undefined) {
// no on going multipart uploads, return empty listing
return cb(null, {
IsTruncated: false,
NextMarker: undefined,
MaxKeys: 0,
});
}
return getMultipartUploadListing(bucket, listingParams, cb);
});
});
},
};
module.exports = metastore;

View File

@ -0,0 +1,62 @@
# bucket_mem design
## RATIONALE
The bucket API will be used for managing buckets behind the S3 interface.
We plan to have only 2 backends using this interface:
* One production backend
* One debug backend purely in memory
One important remark here is that we don't want an abstraction but a
duck-typing style interface (different classes MemoryBucket and Bucket having
the same methods putObjectMD(), getObjectMD(), etc).
Notes about the memory backend: The backend is currently a simple key/value
store in memory. The functions actually use nextTick() to emulate the future
asynchronous behavior of the production backend.
## BUCKET API
The bucket API is a very simple API with 5 functions:
- putObjectMD(): put metadata for an object in the bucket
- getObjectMD(): get metadata from the bucket
- deleteObjectMD(): delete metadata for an object from the bucket
- deleteBucketMD(): delete a bucket
- getBucketListObjects(): perform the complex bucket listing AWS search
function with various flavors. This function returns a response in a
ListBucketResult object.
getBucketListObjects(prefix, marker, delimiter, maxKeys, callback) behavior is
the following:
prefix (not required): Limits the response to keys that begin with the
specified prefix. You can use prefixes to separate a bucket into different
groupings of keys. (You can think of using prefix to make groups in the same
way you'd use a folder in a file system.)
marker (not required): Specifies the key to start with when listing objects in
a bucket. Amazon S3 returns object keys in alphabetical order, starting with
key after the marker in order.
delimiter (not required): A delimiter is a character you use to group keys.
All keys that contain the same string between the prefix, if specified, and the
first occurrence of the delimiter after the prefix are grouped under a single
result element, CommonPrefixes. If you don't specify the prefix parameter, then
the substring starts at the beginning of the key. The keys that are grouped
under CommonPrefixes are not returned elsewhere in the response.
maxKeys: Sets the maximum number of keys returned in the response body. You can
add this to your request if you want to retrieve fewer than the default 1000
keys. The response might contain fewer keys but will never contain more. If
there are additional keys that satisfy the search criteria but were not
returned because maxKeys was exceeded, the response contains an attribute of
IsTruncated set to true and a NextMarker. To return the additional keys, call
the function again using NextMarker as your marker argument in the function.
Any key that does not contain the delimiter will be returned individually in
Contents rather than in CommonPrefixes.
If there is an error, the error subfield is returned in the response.

View File

@ -0,0 +1,51 @@
function markerFilterMPU(allMarkers, array) {
const { keyMarker, uploadIdMarker } = allMarkers;
for (let i = 0; i < array.length; i++) {
// If the keyMarker is the same as the key,
// check the uploadIdMarker. If uploadIdMarker is the same
// as or alphabetically after the uploadId of the item,
// eliminate the item.
if (uploadIdMarker && keyMarker === array[i].key) {
const laterId =
[uploadIdMarker, array[i].uploadId].sort()[1];
if (array[i].uploadId === laterId) {
break;
} else {
array.shift();
i--;
}
} else {
// If the keyMarker is alphabetically after the key
// of the item in the array, eliminate the item from the array.
const laterItem =
[keyMarker, array[i].key].sort()[1];
if (keyMarker === array[i].key || keyMarker === laterItem) {
array.shift();
i--;
} else {
break;
}
}
}
return array;
}
function prefixFilter(prefix, array) {
for (let i = 0; i < array.length; i++) {
if (array[i].indexOf(prefix) !== 0) {
array.splice(i, 1);
i--;
}
}
return array;
}
function isKeyInContents(responseObject, key) {
return responseObject.Contents.some(val => val.key === key);
}
module.exports = {
markerFilterMPU,
prefixFilter,
isKeyInContents,
};

View File

@ -0,0 +1,148 @@
const { errors } = require('arsenal');
const { markerFilterMPU, prefixFilter } = require('./bucket_utilities');
const { ListMultipartUploadsResult } = require('./ListMultipartUploadsResult');
const { metadata } = require('./metadata');
const defaultMaxKeys = 1000;
function getMultipartUploadListing(bucket, params, callback) {
const { delimiter, keyMarker,
uploadIdMarker, prefix, queryPrefixLength, splitter } = params;
const splitterLen = splitter.length;
const maxKeys = params.maxKeys !== undefined ?
Number.parseInt(params.maxKeys, 10) : defaultMaxKeys;
const response = new ListMultipartUploadsResult();
const keyMap = metadata.keyMaps.get(bucket.getName());
if (prefix) {
response.Prefix = prefix;
if (typeof prefix !== 'string') {
return callback(errors.InvalidArgument);
}
}
if (keyMarker) {
response.KeyMarker = keyMarker;
if (typeof keyMarker !== 'string') {
return callback(errors.InvalidArgument);
}
}
if (uploadIdMarker) {
response.UploadIdMarker = uploadIdMarker;
if (typeof uploadIdMarker !== 'string') {
return callback(errors.InvalidArgument);
}
}
if (delimiter) {
response.Delimiter = delimiter;
if (typeof delimiter !== 'string') {
return callback(errors.InvalidArgument);
}
}
if (maxKeys && typeof maxKeys !== 'number') {
return callback(errors.InvalidArgument);
}
// Sort uploads alphatebetically by objectKey and if same objectKey,
// then sort in ascending order by time initiated
let uploads = [];
keyMap.forEach((val, key) => {
uploads.push(key);
});
uploads.sort((a, b) => {
const aIndex = a.indexOf(splitter);
const bIndex = b.indexOf(splitter);
const aObjectKey = a.substring(aIndex + splitterLen);
const bObjectKey = b.substring(bIndex + splitterLen);
const aInitiated = keyMap.get(a).initiated;
const bInitiated = keyMap.get(b).initiated;
if (aObjectKey === bObjectKey) {
if (Date.parse(aInitiated) >= Date.parse(bInitiated)) {
return 1;
}
if (Date.parse(aInitiated) < Date.parse(bInitiated)) {
return -1;
}
}
return (aObjectKey < bObjectKey) ? -1 : 1;
});
// Edit the uploads array so it only
// contains keys that contain the prefix
uploads = prefixFilter(prefix, uploads);
uploads = uploads.map(stringKey => {
const index = stringKey.indexOf(splitter);
const index2 = stringKey.indexOf(splitter, index + splitterLen);
const storedMD = keyMap.get(stringKey);
return {
key: stringKey.substring(index + splitterLen, index2),
uploadId: stringKey.substring(index2 + splitterLen),
bucket: storedMD.eventualStorageBucket,
initiatorID: storedMD.initiator.ID,
initiatorDisplayName: storedMD.initiator.DisplayName,
ownerID: storedMD['owner-id'],
ownerDisplayName: storedMD['owner-display-name'],
storageClass: storedMD['x-amz-storage-class'],
initiated: storedMD.initiated,
};
});
// If keyMarker specified, edit the uploads array so it
// only contains keys that occur alphabetically after the marker.
// If there is also an uploadIdMarker specified, filter to eliminate
// any uploads that share the keyMarker and have an uploadId before
// the uploadIdMarker.
if (keyMarker) {
const allMarkers = {
keyMarker,
uploadIdMarker,
};
uploads = markerFilterMPU(allMarkers, uploads);
}
// Iterate through uploads and filter uploads
// with keys containing delimiter
// into response.CommonPrefixes and filter remaining uploads
// into response.Uploads
for (let i = 0; i < uploads.length; i++) {
const currentUpload = uploads[i];
// If hit maxKeys, stop adding keys to response
if (response.MaxKeys >= maxKeys) {
response.IsTruncated = true;
break;
}
// If a delimiter is specified, find its
// index in the current key AFTER THE OCCURRENCE OF THE PREFIX
// THAT WAS SENT IN THE QUERY (not the prefix including the splitter
// and other elements)
let delimiterIndexAfterPrefix = -1;
const currentKeyWithoutPrefix =
currentUpload.key.slice(queryPrefixLength);
let sliceEnd;
if (delimiter) {
delimiterIndexAfterPrefix = currentKeyWithoutPrefix
.indexOf(delimiter);
sliceEnd = delimiterIndexAfterPrefix + queryPrefixLength;
}
// If delimiter occurs in current key, add key to
// response.CommonPrefixes.
// Otherwise add upload to response.Uploads
if (delimiterIndexAfterPrefix > -1) {
const keySubstring = currentUpload.key.slice(0, sliceEnd + 1);
response.addCommonPrefix(keySubstring);
} else {
response.NextKeyMarker = currentUpload.key;
response.NextUploadIdMarker = currentUpload.uploadId;
response.addUpload(currentUpload);
}
}
// `response.MaxKeys` should be the value from the original `MaxUploads`
// parameter specified by the user (or else the default 1000). Redefine it
// here, so it does not equal the value of `uploads.length`.
response.MaxKeys = maxKeys;
// If `response.MaxKeys` is 0, `response.IsTruncated` should be `false`.
response.IsTruncated = maxKeys === 0 ? false : response.IsTruncated;
return callback(null, response);
}
module.exports = getMultipartUploadListing;

View File

@ -0,0 +1,8 @@
const metadata = {
buckets: new Map,
keyMaps: new Map,
};
module.exports = {
metadata,
};

View File

@ -0,0 +1,274 @@
const async = require('async');
const { errors } = require('arsenal');
const metadata = require('./wrapper');
const BucketInfo = require('arsenal').models.BucketInfo;
const { isBucketAuthorized, isObjAuthorized } =
require('../api/apiUtils/authorization/aclChecks');
const bucketShield = require('../api/apiUtils/bucket/bucketShield');
/** _parseListEntries - parse the values returned in a listing by metadata
* @param {object[]} entries - Version or Content entries in a metadata listing
* @param {string} entries[].key - metadata key
* @param {string} entries[].value - stringified object metadata
* @return {object} - mapped array with parsed value or JSON parsing err
*/
function _parseListEntries(entries) {
return entries.map(entry => {
if (typeof entry.value === 'string') {
const tmp = JSON.parse(entry.value);
return {
key: entry.key,
value: {
Size: tmp['content-length'],
ETag: tmp['content-md5'],
VersionId: tmp.versionId,
IsNull: tmp.isNull,
IsDeleteMarker: tmp.isDeleteMarker,
LastModified: tmp['last-modified'],
Owner: {
DisplayName: tmp['owner-display-name'],
ID: tmp['owner-id'],
},
StorageClass: tmp['x-amz-storage-class'],
// MPU listing properties
Initiated: tmp.initiated,
Initiator: tmp.initiator,
EventualStorageBucket: tmp.eventualStorageBucket,
partLocations: tmp.partLocations,
creationDate: tmp.creationDate,
},
};
}
return entry;
});
}
/** parseListEntries - parse the values returned in a listing by metadata
* @param {object[]} entries - Version or Content entries in a metadata listing
* @param {string} entries[].key - metadata key
* @param {string} entries[].value - stringified object metadata
* @return {(object|Error)} - mapped array with parsed value or JSON parsing err
*/
function parseListEntries(entries) {
// wrap private function in a try/catch clause
// just in case JSON parsing throws an exception
try {
return _parseListEntries(entries);
} catch (e) {
return e;
}
}
/** getNullVersion - return metadata of null version if it exists
* @param {object} objMD - metadata of master version
* @param {string} bucketName - name of bucket
* @param {string} objectKey - name of object key
* @param {RequestLogger} log - request logger
* @param {function} cb - callback
* @return {undefined} - and call callback with params err, objMD of null ver
*/
function getNullVersion(objMD, bucketName, objectKey, log, cb) {
const options = {};
if (objMD.isNull || !objMD.versionId) {
// null version is current version
log.debug('found null version');
return process.nextTick(() => cb(null, objMD));
}
if (objMD.nullVersionId) {
// the latest version is not the null version, but null version exists
log.debug('null version exists, get the null version');
options.versionId = objMD.nullVersionId;
return metadata.getObjectMD(bucketName, objectKey, options, log, cb);
}
log.debug('could not find a null version');
return process.nextTick(() => cb());
}
/** metadataGetBucketAndObject - retrieves bucket and specified version
* NOTE: If the value of `versionId` param is 'null', this function returns the
* master version objMD. The null version object md must be retrieved in a
* separate step using the master object md: see getNullVersion().
* @param {string} requestType - type of request
* @param {string} bucketName - name of bucket
* @param {string} objectKey - name of object key
* @param {string} [versionId] - version of object to retrieve
* @param {RequestLogger} log - request logger
* @param {function} cb - callback
* @return {undefined} - and call callback with err, bucket md and object md
*/
function metadataGetBucketAndObject(requestType, bucketName, objectKey,
versionId, log, cb) {
const options = {
// if attempting to get 'null' version, must retrieve null version id
// from most current object md (versionId = undefined)
versionId: versionId === 'null' ? undefined : versionId,
};
return metadata.getBucketAndObjectMD(bucketName, objectKey, options, log,
(err, data) => {
if (err) {
log.debug('metadata get failed', { error: err });
return cb(err);
}
const bucket = data.bucket ? BucketInfo.deSerialize(data.bucket) :
undefined;
const obj = data.obj ? JSON.parse(data.obj) : undefined;
if (!bucket) {
log.debug('bucketAttrs is undefined', {
bucket: bucketName,
method: 'metadataGetBucketAndObject',
});
return cb(errors.NoSuchBucket);
}
if (bucketShield(bucket, requestType)) {
log.debug('bucket is shielded from request', {
requestType,
method: 'metadataGetBucketAndObject',
});
return cb(errors.NoSuchBucket);
}
log.trace('found bucket in metadata');
return cb(null, bucket, obj);
});
}
/** metadataGetObject - retrieves specified object or version from metadata
* @param {string} bucketName - name of bucket
* @param {string} objectKey - name of object key
* @param {string} [versionId] - version of object to retrieve
* @param {RequestLogger} log - request logger
* @param {function} cb - callback
* @return {undefined} - and call callback with err, bucket md and object md
*/
function metadataGetObject(bucketName, objectKey, versionId, log, cb) {
const options = {
// if attempting to get 'null' version, must first retrieve null version
// id from most current object md (by setting versionId as undefined
// we retrieve the most current object md)
versionId: versionId === 'null' ? undefined : versionId,
};
return metadata.getObjectMD(bucketName, objectKey, options, log,
(err, objMD) => {
if (err) {
log.debug('err getting object MD from metadata',
{ error: err });
return cb(err);
}
if (versionId === 'null') {
return getNullVersion(objMD, bucketName, objectKey, log, cb);
}
return cb(null, objMD);
});
}
/** metadataValidateBucketAndObj - retrieve bucket and object md from metadata
* and check if user is authorized to access them.
* @param {object} params - function parameters
* @param {AuthInfo} params.authInfo - AuthInfo class instance, requester's info
* @param {string} params.bucketName - name of bucket
* @param {string} params.objectKey - name of object
* @param {string} [params.versionId] - version id if getting specific version
* @param {string} params.requestType - type of request
* @param {RequestLogger} log - request logger
* @param {function} callback - callback
* @return {undefined} - and call callback with params err, bucket md
*/
function metadataValidateBucketAndObj(params, log, callback) {
const { authInfo, bucketName, objectKey, versionId, requestType } = params;
const canonicalID = authInfo.getCanonicalID();
async.waterfall([
function getBucketAndObjectMD(next) {
return metadataGetBucketAndObject(requestType, bucketName,
objectKey, versionId, log, next);
},
function checkBucketAuth(bucket, objMD, next) {
if (!isBucketAuthorized(bucket, requestType, canonicalID)) {
log.debug('access denied for user on bucket', { requestType });
return next(errors.AccessDenied, bucket);
}
return next(null, bucket, objMD);
},
function handleNullVersionGet(bucket, objMD, next) {
if (objMD && versionId === 'null') {
return getNullVersion(objMD, bucketName, objectKey, log,
(err, nullVer) => next(err, bucket, nullVer));
}
return next(null, bucket, objMD);
},
function checkObjectAuth(bucket, objMD, next) {
if (!objMD) {
return next(null, bucket);
}
if (!isObjAuthorized(bucket, objMD, requestType, canonicalID)) {
log.debug('access denied for user on object', { requestType });
return next(errors.AccessDenied, bucket);
}
return next(null, bucket, objMD);
},
], (err, bucket, objMD) => {
if (err) {
// still return bucket for cors headers
return callback(err, bucket);
}
return callback(null, bucket, objMD);
});
}
/** metadataGetBucket - retrieves bucket from metadata, returning error if
* bucket is shielded
* @param {string} requestType - type of request
* @param {string} bucketName - name of bucket
* @param {RequestLogger} log - request logger
* @param {function} cb - callback
* @return {undefined} - and call callback with err, bucket md
*/
function metadataGetBucket(requestType, bucketName, log, cb) {
return metadata.getBucket(bucketName, log, (err, bucket) => {
if (err) {
log.debug('metadata getbucket failed', { error: err });
return cb(err);
}
if (bucketShield(bucket, requestType)) {
log.debug('bucket is shielded from request', {
requestType,
method: 'metadataGetBucketAndObject',
});
return cb(errors.NoSuchBucket);
}
log.trace('found bucket in metadata');
return cb(null, bucket);
});
}
/** metadataValidateBucket - retrieve bucket from metadata and check if user
* is authorized to access it
* @param {object} params - function parameters
* @param {AuthInfo} params.authInfo - AuthInfo class instance, requester's info
* @param {string} params.bucketName - name of bucket
* @param {string} params.requestType - type of request
* @param {RequestLogger} log - request logger
* @param {function} callback - callback
* @return {undefined} - and call callback with params err, bucket md
*/
function metadataValidateBucket(params, log, callback) {
const { authInfo, bucketName, requestType } = params;
const canonicalID = authInfo.getCanonicalID();
return metadataGetBucket(requestType, bucketName, log, (err, bucket) => {
if (err) {
return callback(err);
}
// still return bucket for cors headers
if (!isBucketAuthorized(bucket, requestType, canonicalID)) {
log.debug('access denied for user on bucket', { requestType });
return callback(errors.AccessDenied, bucket);
}
return callback(null, bucket);
});
}
module.exports = {
parseListEntries,
metadataGetObject,
metadataValidateBucketAndObj,
metadataValidateBucket,
};

View File

@ -0,0 +1,173 @@
# Mongoclient
We introduce a new metadata backend called *mongoclient* for
[MongoDB](https://www.mongodb.com). This backend takes advantage of
MongoDB being a document store to store the metadata (bucket and
object attributes) as JSON objects.
## Overall Design
The mongoclient backend strictly follows the metadata interface that
stores bucket and object attributes, which consists of the methods
createBucket(), getBucketAttributes(), getBucketAndObject()
(attributes), putBucketAttributes(), deleteBucket(), putObject(),
getObject(), deleteObject(), listObject(), listMultipartUploads() and
the management methods getUUID(), getDiskUsage() and countItems(). The
mongoclient backend also knows how to deal with versioning, it is also
compatible with the various listing algorithms implemented in Arsenal.
FIXME: There should be a document describing the metadata (currently
duck-typing) interface.
### Why Using MongoDB for Storing Bucket and Object Attributes
We chose MongoDB for various reasons:
- MongoDB supports replication, especially through the Raft protocol.
- MongoDB supports a basic replication scheme called 'Replica Set' and
more advanced sharding schemes if required.
- MongoDB is open source and an enterprise standard.
- MongoDB is a document store (natively supports JSON) and supports a
very flexible search interface.
### Choice of Mongo Client Library
We chose to use the official MongoDB driver for NodeJS:
[https://github.com/mongodb/node-mongodb-native](https://github.com/mongodb/node-mongodb-native)
### Granularity for Buckets
We chose to have one collection for one bucket mapping. First because
in a simple mode of replication called 'replica set' it works from the
get-go, but if one or many buckets grow to big it is possible to use
more advanced schemes such as sharding. MongoDB supports a mix of
sharded and non-sharded collections.
### Storing Database Information
We need a special collection called the *Infostore* (stored under the
name __infostore which is impossible to create through the S3 bucket
naming scheme) to store specific database properties such as the
unique *uuid* for Orbit.
### Storing Bucket Attributes
We need to use a special collection called the *Metastore* (stored
under the name __metastore which is impossible to create through the
S3 bucket naming scheme).
### Versioning Format
We chose to keep the same versioning format that we use in some other
Scality products in order to facilitate the compatibility between the
different products.
FIXME: Document the versioning internals in the upper layers and
document the versioning format
### Dealing with Concurrency
We chose not to use transactions (aka
[https://docs.mongodb.com/manual/tutorial/perform-two-phase-commits/)
because it is a known fact there is an overhead of using them, and we
thought there was no real need for them since we could leverage Mongo
ordered operations guarantees and atomic writes.
Example of corner cases:
#### CreateBucket()
Since it is not possible to create a collection AND at the same time
register the bucket in the Metastore we chose to only update the
Metastore. A non-existing collection (NamespaceNotFound error in
Mongo) is one possible normal state for an empty bucket.
#### DeleteBucket()
In this case the bucket is *locked* by the upper layers (use of a
transient delete flag) so we don't have to worry about that and by the
fact the bucket is empty neither (which is also checked by the upper
layers).
We first drop() the collection and then we asynchronously delete the
bucket name entry from the metastore (the removal from the metastore
is atomic which is not absolutely necessary in this case but more
robust in term of design).
If we fail in between we still have an entry in the metastore which is
good because we need to manage the delete flag. For the upper layers
the operation has not completed until this flag is removed. The upper
layers will restart the deleteBucket() which is fine because we manage
the case where the collection does not exist.
#### PutObject() with a Version
We need to store the versioned object then update the master object
(the latest version). For this we use the
[BulkWrite](http://mongodb.github.io/node-mongodb-native/3.0/api/Collection.html#bulkWrite)
method. This is not a transaction but guarantees that the 2 operations
will happen sequentially in the MongoDB oplog. Indeed if the
BulkWrite() fails in between we would end up creating an orphan (which
is not critical) but if the operation succeeds then we are sure that
the master is always pointing to the right object. If there is a
concurrency between 2 clients then we are sure that the 2 groups of
operations will be clearly decided in the oplog (the last writer will
win).
#### DeleteObject()
This is probably the most complex case to manage because it involves a
lot of different cases:
##### Deleting an Object when Versioning is not Enabled
This case is a straightforward atomic delete. Atomicity is not really
required because we assume version IDs are random enough but it is
more robust to do so.
##### Deleting an Object when Versioning is Enabled
This case is more complex since we have to deal with the 2 cases:
Case 1: The caller asks for a deletion of a version which is not a master:
This case is a straight-forward atomic delete.
Case 2: The caller asks for a deletion of a version which is the master: In
this case we need to create a special flag called PHD (as PlaceHolDer)
that indicates the master is no longer valid (with a new unique
virtual version ID). We force the ordering of operations in a
bulkWrite() to first replace the master with the PHD flag and then
physically delete the version. If the call fail in between we will be
left with a master with a PHD flag. If the call succeeds we try to
find if the master with the PHD flag is left alone in such case we
delete it otherwise we trigger an asynchronous repair that will spawn
after AYNC_REPAIR_TIMEOUT=15s that will reassign the master to the
latest version.
In all cases the physical deletion or the repair of the master are
checked against the PHD flag AND the actual unique virtual version
ID. We do this to check against potential concurrent deletions,
repairs or updates. Only the last writer/deleter has the right to
physically perform the operation, otherwise it is superseded by other
operations.
##### Getting an object with a PHD flag
If the caller is asking for the latest version of an object and the
PHD flag is set we perform a search on the bucket to find the latest
version and we return it.
#### Listing Objects
The mongoclient backend implements a readable key/value stream called
*MongoReadStream* that follows the LevelDB duck typing interface used
in Arsenal/lib/algos listing algorithms. Note it does not require any
LevelDB package.
#### Generating the UUID
To avoid race conditions we always (try to) generate a new UUID and we
condition the insertion to the non-existence of the document.

View File

@ -0,0 +1,766 @@
/*
* we assume good default setting of write concern is good for all
* bulk writes. Note that bulk writes are not transactions but ordered
* writes. They may fail in between. To some extend those situations
* may generate orphans but not alter the proper conduct of operations
* (what he user wants and what we acknowledge to the user).
*
* Orphan situations may be recovered by the Lifecycle.
*
* We use proper atomic operations when needed.
*/
const async = require('async');
const arsenal = require('arsenal');
const logger = require('../../utilities/logger');
const constants = require('../../../constants');
const { config } = require('../../Config');
const errors = arsenal.errors;
const versioning = arsenal.versioning;
const BucketInfo = arsenal.models.BucketInfo;
const MongoClient = require('mongodb').MongoClient;
const Uuid = require('uuid');
const diskusage = require('diskusage');
const genVID = versioning.VersionID.generateVersionId;
const MongoReadStream = require('./readStream');
const MongoUtils = require('./utils');
const USERSBUCKET = '__usersbucket';
const METASTORE = '__metastore';
const INFOSTORE = '__infostore';
const __UUID = 'uuid';
const ASYNC_REPAIR_TIMEOUT = 15000;
let uidCounter = 0;
const VID_SEP = versioning.VersioningConstants.VersionId.Separator;
function generateVersionId() {
// generate a unique number for each member of the nodejs cluster
return genVID(`${process.pid}.${uidCounter++}`,
config.replicationGroupId);
}
function formatVersionKey(key, versionId) {
return `${key}${VID_SEP}${versionId}`;
}
function inc(str) {
return str ? (str.slice(0, str.length - 1) +
String.fromCharCode(str.charCodeAt(str.length - 1) + 1)) : str;
}
const VID_SEPPLUS = inc(VID_SEP);
function generatePHDVersion(versionId) {
return {
isPHD: true,
versionId,
};
}
class MongoClientInterface {
constructor() {
const mongoUrl =
`mongodb://${config.mongodb.host}:${config.mongodb.port}`;
this.logger = logger;
this.client = null;
this.db = null;
this.logger.debug(`connecting to ${mongoUrl}`);
// FIXME: constructors shall not have side effect so there
// should be an async_init(cb) method in the wrapper to
// initialize this backend
MongoClient.connect(mongoUrl, (err, client) => {
if (err) {
throw (errors.InternalError);
}
this.logger.debug('connected to mongodb');
this.client = client;
this.db = client.db(config.mongodb.database, {
ignoreUndefined: true,
});
this.usersBucketHack();
});
}
usersBucketHack() {
/* FIXME: Since the bucket creation API is expecting the
usersBucket to have attributes, we pre-create the
usersBucket attributes here (see bucketCreation.js line
36)*/
const usersBucketAttr = new BucketInfo(constants.usersBucket,
'admin', 'admin', new Date().toJSON(),
BucketInfo.currentModelVersion());
this.createBucket(
constants.usersBucket,
usersBucketAttr, {}, err => {
if (err) {
this.logger.fatal('error writing usersBucket ' +
'attributes to metastore',
{ error: err });
throw (errors.InternalError);
}
});
}
getCollection(name) {
/* mongo has a problem with .. in collection names */
const newName = (name === constants.usersBucket) ?
USERSBUCKET : name;
return this.db.collection(newName);
}
createBucket(bucketName, bucketMD, log, cb) {
// FIXME: there should be a version of BucketInfo.serialize()
// that does not JSON.stringify()
const bucketInfo = BucketInfo.fromObj(bucketMD);
const bucketMDStr = bucketInfo.serialize();
const newBucketMD = JSON.parse(bucketMDStr);
const m = this.getCollection(METASTORE);
// we don't have to test bucket existence here as it is done
// on the upper layers
m.update({
_id: bucketName,
}, {
_id: bucketName,
value: newBucketMD,
}, {
upsert: true,
}, () => cb());
}
getBucketAttributes(bucketName, log, cb) {
const m = this.getCollection(METASTORE);
m.findOne({
_id: bucketName,
}, {}, (err, doc) => {
if (err) {
return cb(errors.InternalError);
}
if (!doc) {
return cb(errors.NoSuchBucket);
}
// FIXME: there should be a version of BucketInfo.deserialize()
// that properly inits w/o JSON.parse()
const bucketMDStr = JSON.stringify(doc.value);
const bucketMD = BucketInfo.deSerialize(bucketMDStr);
return cb(null, bucketMD);
});
}
getBucketAndObject(bucketName, objName, params, log, cb) {
this.getBucketAttributes(bucketName, log, (err, bucket) => {
if (err) {
return cb(err);
}
this.getObject(bucketName, objName, params, log, (err, obj) => {
if (err) {
if (err === errors.NoSuchKey) {
return cb(null,
{ bucket:
BucketInfo.fromObj(bucket).serialize(),
});
}
return cb(err);
}
return cb(null, {
bucket: BucketInfo.fromObj(bucket).serialize(),
obj: JSON.stringify(obj),
});
});
return undefined;
});
}
putBucketAttributes(bucketName, bucketMD, log, cb) {
// FIXME: there should be a version of BucketInfo.serialize()
// that does not JSON.stringify()
const bucketInfo = BucketInfo.fromObj(bucketMD);
const bucketMDStr = bucketInfo.serialize();
const newBucketMD = JSON.parse(bucketMDStr);
const m = this.getCollection(METASTORE);
m.update({
_id: bucketName,
}, {
_id: bucketName,
value: newBucketMD,
}, {
upsert: true,
}, () => cb());
}
/*
* Delete bucket from metastore
*/
deleteBucketStep2(bucketName, log, cb) {
const m = this.getCollection(METASTORE);
m.findOneAndDelete({
_id: bucketName,
}, {}, (err, result) => {
if (err) {
return cb(errors.InternalError);
}
if (result.ok !== 1) {
return cb(errors.InternalError);
}
return cb(null);
});
}
/*
* Drop the bucket then process to step 2. Checking
* the count is already done by the upper layer. We don't need to be
* atomic because the call is protected by a delete_pending flag
* in the upper layer.
* 2 cases here:
* 1) the collection may not yet exist (or being already dropped
* by a previous call)
* 2) the collection may exist.
*/
deleteBucket(bucketName, log, cb) {
const c = this.getCollection(bucketName);
c.drop({}, err => {
if (err) {
if (err.codeName === 'NamespaceNotFound') {
return this.deleteBucketStep2(bucketName, log, cb);
}
return cb(errors.InternalError);
}
return this.deleteBucketStep2(bucketName, log, cb);
});
}
/*
* In this case we generate a versionId and
* sequentially create the object THEN update the master
*/
putObjectVerCase1(c, bucketName, objName, objVal, params, log, cb) {
const versionId = generateVersionId();
// eslint-disable-next-line
objVal.versionId = versionId;
const vObjName = formatVersionKey(objName, versionId);
c.bulkWrite([{
updateOne: {
filter: {
_id: vObjName,
},
update: {
_id: vObjName, value: objVal,
},
upsert: true,
},
}, {
updateOne: {
filter: {
_id: objName,
},
update: {
_id: objName, value: objVal,
},
upsert: true,
},
}], {
ordered: 1,
}, () => cb(null, `{"versionId": "${versionId}"}`));
}
/*
* Case used when versioning has been disabled after objects
* have been created with versions
*/
putObjectVerCase2(c, bucketName, objName, objVal, params, log, cb) {
const versionId = generateVersionId();
// eslint-disable-next-line
objVal.versionId = versionId;
c.update({
_id: objName,
}, {
_id: objName,
value: objVal,
}, {
upsert: true,
}, () => cb(null, `{"versionId": "${objVal.versionId}"}`));
}
/*
* In this case the aller provides a versionId. This function will
* sequentially update the object with given versionId THEN the
* master iff the provided versionId matches the one of the master
*/
putObjectVerCase3(c, bucketName, objName, objVal, params, log, cb) {
// eslint-disable-next-line
objVal.versionId = params.versionId;
const vObjName = formatVersionKey(objName, params.versionId);
c.bulkWrite([{
updateOne: {
filter: {
_id: vObjName,
},
update: {
_id: vObjName, value: objVal,
},
upsert: true,
},
}, {
updateOne: {
// eslint-disable-next-line
filter: {
_id: objName,
'value.versionId': params.versionId,
},
update: {
_id: objName, value: objVal,
},
upsert: true,
},
}], {
ordered: 1,
}, () => cb(null, `{"versionId": "${objVal.versionId}"}`));
}
/*
* Put object when versioning is not enabled
*/
putObjectNoVer(c, bucketName, objName, objVal, params, log, cb) {
c.update({
_id: objName,
}, {
_id: objName,
value: objVal,
}, {
upsert: true,
}, () => cb());
}
putObject(bucketName, objName, objVal, params, log, cb) {
MongoUtils.serialize(objVal);
const c = this.getCollection(bucketName);
if (params && params.versioning && !params.versionId) {
return this.putObjectVerCase1(c, bucketName, objName, objVal,
params, log, cb);
} else if (params && params.versionId === '') {
return this.putObjectVerCase2(c, bucketName, objName, objVal,
params, log, cb);
} else if (params && params.versionId) {
return this.putObjectVerCase3(c, bucketName, objName, objVal,
params, log, cb);
}
return this.putObjectNoVer(c, bucketName, objName, objVal,
params, log, cb);
}
getObject(bucketName, objName, params, log, cb) {
const c = this.getCollection(bucketName);
if (params && params.versionId) {
// eslint-disable-next-line
objName = formatVersionKey(objName, params.versionId);
}
c.findOne({
_id: objName,
}, {}, (err, doc) => {
if (err) {
return cb(errors.InternalError);
}
if (!doc) {
return cb(errors.NoSuchKey);
}
if (doc.value.isPHD) {
this.getLatestVersion(c, objName, log, (err, value) => {
if (err) {
log.error('getting latest version', err);
return cb(err);
}
return cb(null, value);
});
return undefined;
}
MongoUtils.unserialize(doc.value);
return cb(null, doc.value);
});
}
/*
* This function return the latest version
*/
getLatestVersion(c, objName, log, cb) {
c.find({
_id: {
$gt: objName,
$lt: `${objName}${VID_SEPPLUS}`,
},
}, {}).
sort({
_id: 1,
}).
limit(1).
toArray(
(err, keys) => {
if (err) {
return cb(errors.InternalError);
}
if (keys.length === 0) {
return cb(errors.NoSuchKey);
}
MongoUtils.unserialize(keys[0].value);
return cb(null, keys[0].value);
});
}
/*
* repair the master with a new value. There can be
* race-conditions or legit updates so place an atomic condition
* on PHD flag and mst version.
*/
repair(c, objName, objVal, mst, log, cb) {
MongoUtils.serialize(objVal);
// eslint-disable-next-line
c.findOneAndReplace({
_id: objName,
'value.isPHD': true,
'value.versionId': mst.versionId,
}, {
_id: objName,
value: objVal,
}, {
upsert: true,
}, (err, result) => {
if (err) {
return cb(errors.InternalError);
}
if (result.ok !== 1) {
return cb(errors.InternalError);
}
return cb(null);
});
}
/*
* Get the latest version and repair. The process is safe because
* we never replace a non-PHD master
*/
asyncRepair(c, objName, mst, log) {
this.getLatestVersion(c, objName, log, (err, value) => {
if (err) {
log.error('async-repair: getting latest version', err);
return undefined;
}
this.repair(c, objName, value, mst, log, err => {
if (err) {
log.error('async-repair failed', err);
return undefined;
}
log.debug('async-repair success');
return undefined;
});
return undefined;
});
}
/*
* the master is a PHD so we try to see if it is the latest of its
* kind to get rid of it, otherwise we asynchronously repair it
*/
deleteOrRepairPHD(c, bucketName, objName, mst, log, cb) {
// Check if there are other versions available
this.getLatestVersion(c, objName, log, err => {
if (err) {
if (err === errors.NoSuchKey) {
// We try to delete the master. A race condition
// is possible here: another process may recreate
// a master or re-delete it in between so place an
// atomic condition on the PHD flag and the mst
// version:
// eslint-disable-next-line
c.findOneAndDelete({
_id: objName,
'value.isPHD': true,
'value.versionId': mst.versionId,
}, {}, err => {
if (err) {
return cb(errors.InternalError);
}
// do not test result.ok === 1 because
// both cases are expected
return cb(null);
});
return undefined;
}
return cb(err);
}
// We have other versions available so repair:
setTimeout(() => {
this.asyncRepair(c, objName, mst, log);
}, ASYNC_REPAIR_TIMEOUT);
return cb(null);
});
}
/*
* Delete object when versioning is enabled and the version is
* master. In this case we sequentially update the master with a
* PHD flag (placeholder) and a unique non-existing version THEN
* we delete the specified versioned object. THEN we try to delete
* or repair the PHD we just created
*/
deleteObjectVerMaster(c, bucketName, objName, params, log, cb) {
const vObjName = formatVersionKey(objName, params.versionId);
const _vid = generateVersionId();
const mst = generatePHDVersion(_vid);
c.bulkWrite([{
updateOne: {
filter: {
_id: objName,
},
update: {
_id: objName, value: mst,
},
upsert: true,
},
}, {
deleteOne: {
filter: {
_id: vObjName,
},
},
}], {
ordered: 1,
}, () => this.deleteOrRepairPHD(c, bucketName, objName, mst, log, cb));
}
/*
* Delete object when versioning is enabled and the version is
* not master. It is a straight-forward atomic delete
*/
deleteObjectVerNotMaster(c, bucketName, objName, params, log, cb) {
const vObjName = formatVersionKey(objName, params.versionId);
c.findOneAndDelete({
_id: vObjName,
}, {}, (err, result) => {
if (err) {
return cb(errors.InternalError);
}
if (result.ok !== 1) {
return cb(errors.InternalError);
}
return cb(null);
});
}
/*
* Delete object when versioning is enabled. We first find the
* master, if it is already a PHD we have a special processing,
* then we check if it matches the master versionId in such case
* we will create a PHD, otherwise we delete it
*/
deleteObjectVer(c, bucketName, objName, params, log, cb) {
c.findOne({
_id: objName,
}, {}, (err, mst) => {
if (err) {
return cb(errors.InternalError);
}
if (!mst) {
return cb(errors.NoSuchKey);
}
if (mst.value.isPHD ||
mst.value.versionId === params.versionId) {
return this.deleteObjectVerMaster(c, bucketName, objName,
params, log, cb);
}
return this.deleteObjectVerNotMaster(c, bucketName, objName,
params, log, cb);
});
}
/*
* Atomically delete an object when versioning is not enabled
*/
deleteObjectNoVer(c, bucketName, objName, params, log, cb) {
c.findOneAndDelete({
_id: objName,
}, {}, (err, result) => {
if (err) {
return cb(errors.InternalError);
}
if (result.ok !== 1) {
return cb(errors.InternalError);
}
return cb(null);
});
}
deleteObject(bucketName, objName, params, log, cb) {
const c = this.getCollection(bucketName);
if (params && params.versionId) {
return this.deleteObjectVer(c, bucketName, objName,
params, log, cb);
}
return this.deleteObjectNoVer(c, bucketName, objName,
params, log, cb);
}
internalListObject(bucketName, params, log, cb) {
const extName = params.listingType;
const extension = new arsenal.algorithms.list[extName](params, log);
const requestParams = extension.genMDParams();
const c = this.getCollection(bucketName);
let cbDone = false;
const stream = new MongoReadStream(c, requestParams);
stream
.on('data', e => {
if (extension.filter(e) < 0) {
stream.emit('end');
stream.destroy();
}
})
.on('error', err => {
if (!cbDone) {
cbDone = true;
const logObj = {
rawError: err,
error: err.message,
errorStack: err.stack,
};
log.error('error listing objects', logObj);
cb(errors.InternalError);
}
})
.on('end', () => {
if (!cbDone) {
cbDone = true;
const data = extension.result();
cb(null, data);
}
});
return undefined;
}
listObject(bucketName, params, log, cb) {
return this.internalListObject(bucketName, params, log, cb);
}
listMultipartUploads(bucketName, params, log, cb) {
return this.internalListObject(bucketName, params, log, cb);
}
readUUID(log, cb) {
const i = this.getCollection(INFOSTORE);
i.findOne({
_id: __UUID,
}, {}, (err, doc) => {
if (err) {
return cb(errors.InternalError);
}
if (!doc) {
return cb(errors.NoSuchKey);
}
return cb(null, doc.value);
});
}
writeUUIDIfNotExists(uuid, log, cb) {
const i = this.getCollection(INFOSTORE);
i.insert({
_id: __UUID,
value: uuid,
}, {}, err => {
if (err) {
if (err.code === 11000) {
// duplicate key error
// FIXME: define a KeyAlreadyExists error in Arsenal
return cb(errors.EntityAlreadyExists);
}
return cb(errors.InternalError);
}
// FIXME: shoud we check for result.ok === 1 ?
return cb(null);
});
}
/*
* we always try to generate a new UUID in order to be atomic in
* case of concurrency. The write will fail if it already exists.
*/
getUUID(log, cb) {
const uuid = Uuid.v4();
this.writeUUIDIfNotExists(uuid, log, err => {
if (err) {
if (err === errors.InternalError) {
return cb(err);
}
return this.readUUID(log, cb);
}
return cb(null, uuid);
});
}
getDiskUsage(cb) {
// FIXME: for basic one server deployment the infrastructure
// configurator shall set a path to the actual MongoDB volume.
// For Kub/cluster deployments there should be a more sophisticated
// way for guessing free space.
diskusage.check(config.mongodb.path !== undefined ?
config.mongodb.path : '/', cb);
}
countItems(log, cb) {
const res = {
objects: 0,
versions: 0,
buckets: 0,
};
this.db.listCollections().toArray((err, collInfos) => {
async.eachLimit(collInfos, 10, (value, next) => {
if (value.name === METASTORE ||
value.name === INFOSTORE ||
value.name === USERSBUCKET) {
// skip
return next();
}
res.buckets++;
const c = this.getCollection(value.name);
// FIXME: there is currently no way of distinguishing
// master from versions and searching for VID_SEP
// does not work because there cannot be null bytes
// in $regex
c.count({
// eslint-disable-next-line
'value.versionId': {
'$exists': false,
},
}, {}, (err, result) => {
if (err) {
return next(errors.InternalError);
}
res.objects += result;
c.count({
// eslint-disable-next-line
'value.versionId': {
'$exists': true,
},
}, {}, (err, result) => {
if (err) {
return next(errors.InternalError);
}
res.versions += result;
return next();
});
return undefined;
});
return undefined;
}, err => {
if (err) {
return cb(err);
}
return cb(null, res);
});
});
}
}
module.exports = MongoClientInterface;

View File

@ -0,0 +1,133 @@
const Readable = require('stream').Readable;
const MongoUtils = require('./utils');
class MongoReadStream extends Readable {
constructor(c, options) {
super({
objectMode: true,
highWaterMark: 0,
});
if (options.limit === 0) {
return;
}
const query = {
_id: {},
};
if (options.reverse) {
if (options.start) {
query._id.$lte = options.start;
}
if (options.end) {
query._id.$gte = options.end;
}
if (options.gt) {
query._id.$lt = options.gt;
}
if (options.gte) {
query._id.$lte = options.gte;
}
if (options.lt) {
query._id.$gt = options.lt;
}
if (options.lte) {
query._id.$gte = options.lte;
}
} else {
if (options.start) {
query._id.$gte = options.start;
}
if (options.end) {
query._id.$lte = options.end;
}
if (options.gt) {
query._id.$gt = options.gt;
}
if (options.gte) {
query._id.$gte = options.gte;
}
if (options.lt) {
query._id.$lt = options.lt;
}
if (options.lte) {
query._id.$lte = options.lte;
}
}
if (!Object.keys(query._id).length) {
delete query._id;
}
this._cursor = c.find(query).sort({
_id: options.reverse ? -1 : 1,
});
if (options.limit && options.limit !== -1) {
this._cursor = this._cursor.limit(options.limit);
}
this._options = options;
this._destroyed = false;
this.on('end', this._cleanup.bind(this));
}
_read() {
if (this._destroyed) {
return;
}
this._cursor.next((err, doc) => {
if (this._destroyed) {
return;
}
if (err) {
this.emit('error', err);
return;
}
let key = undefined;
let value = undefined;
if (doc) {
key = doc._id;
MongoUtils.unserialize(doc.value);
value = JSON.stringify(doc.value);
}
if (key === undefined && value === undefined) {
this.push(null);
} else if (this._options.keys !== false &&
this._options.values === false) {
this.push(key);
} else if (this._options.keys === false &&
this._options.values !== false) {
this.push(value);
} else {
this.push({
key,
value,
});
}
});
}
_cleanup() {
if (this._destroyed) {
return;
}
this._destroyed = true;
this._cursor.close(err => {
if (err) {
this.emit('error', err);
return;
}
this.emit('close');
});
}
destroy() {
return this._cleanup();
}
}
module.exports = MongoReadStream;

View File

@ -0,0 +1,30 @@
function escape(obj) {
return JSON.parse(JSON.stringify(obj).
replace(/\$/g, '\uFF04').
replace(/\./g, '\uFF0E'));
}
function unescape(obj) {
return JSON.parse(JSON.stringify(obj).
replace(/\uFF04/g, '$').
replace(/\uFF0E/g, '.'));
}
function serialize(objMD) {
// Tags require special handling since dot and dollar are accepted
if (objMD.tags) {
// eslint-disable-next-line
objMD.tags = escape(objMD.tags);
}
}
function unserialize(objMD) {
// Tags require special handling
if (objMD.tags) {
// eslint-disable-next-line
objMD.tags = unescape(objMD.tags);
}
}
module.exports = { escape, unescape, serialize, unserialize };

View File

@ -0,0 +1,256 @@
const errors = require('arsenal').errors;
const BucketClientInterface = require('./bucketclient/backend');
const BucketFileInterface = require('./bucketfile/backend');
const MongoClientInterface = require('./mongoclient/backend');
const BucketInfo = require('arsenal').models.BucketInfo;
const inMemory = require('./in_memory/backend');
const { config } = require('../Config');
let CdmiMetadata;
try {
CdmiMetadata = require('cdmiclient').CdmiMetadata;
} catch (err) {
CdmiMetadata = null;
}
let client;
let implName;
if (config.backends.metadata === 'mem') {
client = inMemory;
implName = 'memorybucket';
} else if (config.backends.metadata === 'file') {
client = new BucketFileInterface();
implName = 'bucketfile';
} else if (config.backends.metadata === 'scality') {
client = new BucketClientInterface();
implName = 'bucketclient';
} else if (config.backends.metadata === 'mongodb') {
client = new MongoClientInterface();
implName = 'mongoclient';
} else if (config.backends.metadata === 'cdmi') {
if (!CdmiMetadata) {
throw new Error('Unauthorized backend');
}
client = new CdmiMetadata({
path: config.cdmi.path,
host: config.cdmi.host,
port: config.cdmi.port,
readonly: config.cdmi.readonly,
});
implName = 'cdmi';
}
const metadata = {
createBucket: (bucketName, bucketMD, log, cb) => {
log.debug('creating bucket in metadata');
client.createBucket(bucketName, bucketMD, log, err => {
if (err) {
log.debug('error from metadata', { implName, error: err });
return cb(err);
}
log.trace('bucket created in metadata');
return cb(err);
});
},
updateBucket: (bucketName, bucketMD, log, cb) => {
log.debug('updating bucket in metadata');
client.putBucketAttributes(bucketName, bucketMD, log, err => {
if (err) {
log.debug('error from metadata', { implName, error: err });
return cb(err);
}
log.trace('bucket updated in metadata');
return cb(err);
});
},
getBucket: (bucketName, log, cb) => {
log.debug('getting bucket from metadata');
client.getBucketAttributes(bucketName, log, (err, data) => {
if (err) {
log.debug('error from metadata', { implName, error: err });
return cb(err);
}
log.trace('bucket retrieved from metadata');
return cb(err, BucketInfo.fromObj(data));
});
},
deleteBucket: (bucketName, log, cb) => {
log.debug('deleting bucket from metadata');
client.deleteBucket(bucketName, log, err => {
if (err) {
log.debug('error from metadata', { implName, error: err });
return cb(err);
}
log.debug('Deleted bucket from Metadata');
return cb(err);
});
},
putObjectMD: (bucketName, objName, objVal, params, log, cb) => {
log.debug('putting object in metadata');
const value = typeof objVal.getValue === 'function' ?
objVal.getValue() : objVal;
client.putObject(bucketName, objName, value, params, log,
(err, data) => {
if (err) {
log.debug('error from metadata', { implName, error: err });
return cb(err);
}
if (data) {
log.debug('object version successfully put in metadata',
{ version: data });
} else {
log.debug('object successfully put in metadata');
}
return cb(err, data);
});
},
getBucketAndObjectMD: (bucketName, objName, params, log, cb) => {
log.debug('getting bucket and object from metadata',
{ database: bucketName, object: objName });
client.getBucketAndObject(bucketName, objName, params, log,
(err, data) => {
if (err) {
log.debug('error from metadata', { implName, err });
return cb(err);
}
log.debug('bucket and object retrieved from metadata',
{ database: bucketName, object: objName });
return cb(err, data);
});
},
getObjectMD: (bucketName, objName, params, log, cb) => {
log.debug('getting object from metadata');
client.getObject(bucketName, objName, params, log, (err, data) => {
if (err) {
log.debug('error from metadata', { implName, err });
return cb(err);
}
log.debug('object retrieved from metadata');
return cb(err, data);
});
},
deleteObjectMD: (bucketName, objName, params, log, cb) => {
log.debug('deleting object from metadata');
client.deleteObject(bucketName, objName, params, log, err => {
if (err) {
log.debug('error from metadata', { implName, err });
return cb(err);
}
log.debug('object deleted from metadata');
return cb(err);
});
},
listObject: (bucketName, listingParams, log, cb) => {
const metadataUtils = require('./metadataUtils');
if (listingParams.listingType === undefined) {
// eslint-disable-next-line
listingParams.listingType = 'Delimiter';
}
client.listObject(bucketName, listingParams, log, (err, data) => {
log.debug('getting object listing from metadata');
if (err) {
log.debug('error from metadata', { implName, err });
return cb(err);
}
log.debug('object listing retrieved from metadata');
if (listingParams.listingType === 'DelimiterVersions') {
// eslint-disable-next-line
data.Versions = metadataUtils.parseListEntries(data.Versions);
if (data.Versions instanceof Error) {
log.error('error parsing metadata listing', {
error: data.Versions,
listingType: listingParams.listingType,
method: 'listObject',
});
return cb(errors.InternalError);
}
return cb(null, data);
}
// eslint-disable-next-line
data.Contents = metadataUtils.parseListEntries(data.Contents);
if (data.Contents instanceof Error) {
log.error('error parsing metadata listing', {
error: data.Contents,
listingType: listingParams.listingType,
method: 'listObject',
});
return cb(errors.InternalError);
}
return cb(null, data);
});
},
listMultipartUploads: (bucketName, listingParams, log, cb) => {
client.listMultipartUploads(bucketName, listingParams, log,
(err, data) => {
log.debug('getting mpu listing from metadata');
if (err) {
log.debug('error from metadata', { implName, err });
return cb(err);
}
log.debug('mpu listing retrieved from metadata');
return cb(err, data);
});
},
switch: newClient => {
client = newClient;
},
getRaftBuckets: (raftId, log, cb) => {
if (!client.getRaftBuckets) {
return cb();
}
return client.getRaftBuckets(raftId, log, cb);
},
checkHealth: (log, cb) => {
if (!client.checkHealth) {
const defResp = {};
defResp[implName] = { code: 200, message: 'OK' };
return cb(null, defResp);
}
return client.checkHealth(implName, log, cb);
},
getUUID: (log, cb) => {
if (!client.getUUID) {
log.debug('returning empty uuid as fallback', { implName });
return cb(null, '');
}
return client.getUUID(log, cb);
},
getDiskUsage: (log, cb) => {
if (!client.getDiskUsage) {
log.debug('returning empty disk usage as fallback', { implName });
return cb(null, {});
}
return client.getDiskUsage(cb);
},
countItems: (log, cb) => {
if (!client.countItems) {
log.debug('returning zero item counts as fallback', { implName });
return cb(null, {
buckets: 0,
objects: 0,
versions: 0,
});
}
return client.countItems(log, cb);
},
};
module.exports = metadata;

View File

@ -1,4 +0,0 @@
#!/usr/bin/env node
'use strict'; // eslint-disable-line strict
require('../lib/nfs/utilities.js').createBucketWithNFSEnabled();

View File

@ -1,108 +0,0 @@
#!/bin/sh
// 2>/dev/null ; exec "$(which nodejs 2>/dev/null || which node)" "$0" "$@"
'use strict'; // eslint-disable-line strict
const { auth } = require('arsenal');
const commander = require('commander');
const http = require('http');
const https = require('https');
const logger = require('../lib/utilities/logger');
function _performSearch(host,
port,
bucketName,
query,
listVersions,
accessKey,
secretKey,
sessionToken,
verbose, ssl) {
const escapedSearch = encodeURIComponent(query);
const options = {
host,
port,
method: 'GET',
path: `/${bucketName}/?search=${escapedSearch}${listVersions ? '&&versions' : ''}`,
headers: {
'Content-Length': 0,
},
rejectUnauthorized: false,
versions: '',
};
if (sessionToken) {
options.headers['x-amz-security-token'] = sessionToken;
}
const transport = ssl ? https : http;
const request = transport.request(options, response => {
if (verbose) {
logger.info('response status code', {
statusCode: response.statusCode,
});
logger.info('response headers', { headers: response.headers });
}
const body = [];
response.setEncoding('utf8');
response.on('data', chunk => body.push(chunk));
response.on('end', () => {
if (response.statusCode >= 200 && response.statusCode < 300) {
logger.info('Success');
process.stdout.write(body.join(''));
process.exit(0);
} else {
logger.error('request failed with HTTP Status ', {
statusCode: response.statusCode,
body: body.join(''),
});
process.exit(1);
}
});
});
// generateV4Headers exepects request object with path that does not
// include query
request.path = `/${bucketName}`;
const requestData = listVersions ? { search: query, versions: '' } : { search: query };
auth.client.generateV4Headers(request, requestData, accessKey, secretKey, 's3');
request.path = `/${bucketName}?search=${escapedSearch}${listVersions ? '&&versions' : ''}`;
if (verbose) {
logger.info('request headers', { headers: request._headers });
}
request.end();
}
/**
* This function is used as a binary to send a request to S3 to perform a
* search on the objects in a bucket
*
* @return {undefined}
*/
function searchBucket() {
// TODO: Include other bucket listing possible query params?
commander
.version('0.0.1')
.option('-a, --access-key <accessKey>', 'Access key id')
.option('-k, --secret-key <secretKey>', 'Secret access key')
.option('-t, --session-token <sessionToken>', 'Session token')
.option('-b, --bucket <bucket>', 'Name of the bucket')
.option('-q, --query <query>', 'Search query')
.option('-h, --host <host>', 'Host of the server')
.option('-p, --port <port>', 'Port of the server')
.option('-s', '--ssl', 'Enable ssl')
.option('-l, --list-versions', 'List all versions of the objects that meet the search query, ' +
'otherwise only list the latest version')
.option('-v, --verbose')
.parse(process.argv);
const { host, port, accessKey, secretKey, sessionToken, bucket, query, listVersions, verbose, ssl } =
commander;
if (!host || !port || !accessKey || !secretKey || !bucket || !query) {
logger.error('missing parameter');
commander.outputHelp();
process.exit(1);
}
_performSearch(host, port, bucket, query, listVersions, accessKey, secretKey, sessionToken, verbose,
ssl);
}
searchBucket();

45
circle.yml Normal file
View File

@ -0,0 +1,45 @@
---
general:
branches:
ignore:
- /^ultron\/.*/ # Ignore ultron/* branches
artifacts:
- coverage/
machine:
node:
version: 6.13.1
services:
- redis
- docker
ruby:
version: "2.4.1"
environment:
CXX: g++-4.9
ENABLE_LOCAL_CACHE: true
REPORT_TOKEN: report-token-1
hosts:
bucketwebsitetester.s3-website-us-east-1.amazonaws.com: 127.0.0.1
dependencies:
override:
- rm -rf node_modules
- npm install
post:
- sudo pip install flake8 yamllint
- sudo pip install s3cmd==1.6.1
# fog and ruby testing dependencies
- gem install fog-aws -v 1.3.0
- gem install mime-types -v 3.1
- gem install rspec -v 3.5
- gem install json
# java sdk dependencies
- sudo apt-get install -y -q default-jdk
test:
override:
- docker run --name squid-proxy -d --net=host
--publish 3128:3128 sameersbn/squid:3.3.8-23
- bash tests.bash:
parallel: true

View File

@ -1,143 +0,0 @@
{
"port": 8000,
"listenOn": [],
"metricsPort": 8002,
"metricsListenOn": [],
"replicationGroupId": "RG001",
"workers": 4,
"restEndpoints": {
"localhost": "us-east-1",
"127.0.0.1": "us-east-1",
"cloudserver-front": "us-east-1",
"s3.docker.test": "us-east-1",
"127.0.0.2": "us-east-1",
"s3.amazonaws.com": "us-east-1",
"zenko-cloudserver-replicator": "us-east-1",
"lb": "us-east-1"
},
"websiteEndpoints": ["s3-website-us-east-1.amazonaws.com",
"s3-website.us-east-2.amazonaws.com",
"s3-website-us-west-1.amazonaws.com",
"s3-website-us-west-2.amazonaws.com",
"s3-website.ap-south-1.amazonaws.com",
"s3-website.ap-northeast-2.amazonaws.com",
"s3-website-ap-southeast-1.amazonaws.com",
"s3-website-ap-southeast-2.amazonaws.com",
"s3-website-ap-northeast-1.amazonaws.com",
"s3-website.eu-central-1.amazonaws.com",
"s3-website-eu-west-1.amazonaws.com",
"s3-website-sa-east-1.amazonaws.com",
"s3-website.localhost",
"s3-website.scality.test",
"zenkoazuretest.blob.core.windows.net"],
"replicationEndpoints": [{
"site": "zenko",
"servers": ["127.0.0.1:8000"],
"default": true
}, {
"site": "us-east-2",
"type": "aws_s3"
}],
"backbeat": {
"host": "localhost",
"port": 8900
},
"workflowEngineOperator": {
"host": "localhost",
"port": 3001
},
"cdmi": {
"host": "localhost",
"port": 81,
"path": "/dewpoint",
"readonly": true
},
"bucketd": {
"bootstrap": ["localhost:9000"]
},
"vaultd": {
"host": "localhost",
"port": 8500
},
"clusters": 1,
"log": {
"logLevel": "info",
"dumpLevel": "error"
},
"healthChecks": {
"allowFrom": ["127.0.0.1/8", "::1"]
},
"metadataClient": {
"host": "localhost",
"port": 9990
},
"dataClient": {
"host": "localhost",
"port": 9991
},
"pfsClient": {
"host": "localhost",
"port": 9992
},
"metadataDaemon": {
"bindAddress": "localhost",
"port": 9990
},
"dataDaemon": {
"bindAddress": "localhost",
"port": 9991
},
"pfsDaemon": {
"bindAddress": "localhost",
"port": 9992
},
"recordLog": {
"enabled": true,
"recordLogName": "s3-recordlog"
},
"mongodb": {
"replicaSetHosts": "localhost:27018,localhost:27019,localhost:27020",
"writeConcern": "majority",
"replicaSet": "rs0",
"readPreference": "primary",
"database": "metadata"
},
"authdata": "authdata.json",
"backends": {
"auth": "file",
"data": "file",
"metadata": "mongodb",
"kms": "file",
"quota": "none"
},
"externalBackends": {
"aws_s3": {
"httpAgent": {
"keepAlive": false,
"keepAliveMsecs": 1000,
"maxFreeSockets": 256,
"maxSockets": null
}
},
"gcp": {
"httpAgent": {
"keepAlive": true,
"keepAliveMsecs": 1000,
"maxFreeSockets": 256,
"maxSockets": null
}
}
},
"requests": {
"viaProxy": false,
"trustedProxyCIDRs": [],
"extractClientIPFromHeader": ""
},
"bucketNotificationDestinations": [
{
"resource": "target1",
"type": "dummy",
"host": "localhost:6000"
}
]
}

View File

@ -1,71 +0,0 @@
{
"port": 8000,
"listenOn": [],
"metricsPort": 8002,
"metricsListenOn": [],
"replicationGroupId": "RG001",
"restEndpoints": {
"localhost": "STANDARD",
"127.0.0.1": "STANDARD",
"yourhostname.ru": "STANDARD"
},
"websiteEndpoints": [
"static.yourhostname.ru"
],
"replicationEndpoints": [ {
"site": "zenko",
"servers": ["127.0.0.1:8000"],
"default": true
} ],
"log": {
"logLevel": "info",
"dumpLevel": "error"
},
"healthChecks": {
"allowFrom": ["127.0.0.1/8", "::1"]
},
"backends": {
"metadata": "mongodb"
},
"mongodb": {
"replicaSetHosts": "127.0.0.1:27017",
"writeConcern": "majority",
"replicaSet": "rs0",
"readPreference": "primary",
"database": "s3",
"authCredentials": {
"username": "s3",
"password": ""
}
},
"externalBackends": {
"aws_s3": {
"httpAgent": {
"keepAlive": false,
"keepAliveMsecs": 1000,
"maxFreeSockets": 256,
"maxSockets": null
}
},
"gcp": {
"httpAgent": {
"keepAlive": true,
"keepAliveMsecs": 1000,
"maxFreeSockets": 256,
"maxSockets": null
}
}
},
"requests": {
"viaProxy": false,
"trustedProxyCIDRs": [],
"extractClientIPFromHeader": ""
},
"bucketNotificationDestinations": [
{
"resource": "target1",
"type": "dummy",
"host": "localhost:6000"
}
]
}

1160
config/Config.js Normal file

File diff suppressed because it is too large Load Diff

23
config/authdata.json Normal file
View File

@ -0,0 +1,23 @@
{
"accounts": [{
"name": "Bart",
"email": "sampleaccount1@sampling.com",
"arn": "arn:aws:iam::123456789012:root",
"canonicalID": "79a59df900b949e55d96a1e698fbacedfd6e09d98eacf8f8d5218e7cd47ef2be",
"shortid": "123456789012",
"keys": [{
"access": "accessKey1",
"secret": "verySecretKey1"
}]
}, {
"name": "Lisa",
"email": "sampleaccount2@sampling.com",
"arn": "arn:aws:iam::123456789013:root",
"canonicalID": "79a59df900b949e55d96a1e698fbacedfd6e09d98eacf8f8d5218e7cd47ef2bf",
"shortid": "123456789013",
"keys": [{
"access": "accessKey2",
"secret": "verySecretKey2"
}]
}]
}

View File

@ -23,8 +23,7 @@
"s3-website-eu-west-1.amazonaws.com",
"s3-website-sa-east-1.amazonaws.com",
"s3-website.localhost",
"s3-website.scality.test",
"zenkoazuretest.blob.core.windows.net"],
"s3-website.scality.test"],
"replicationEndpoints": [{
"site": "zenko",
"servers": ["127.0.0.1:8000"],
@ -40,7 +39,7 @@
"readonly": true
},
"bucketd": {
"bootstrap": ["localhost"]
"bootstrap": ["localhost:9000"]
},
"vaultd": {
"host": "localhost",
@ -75,52 +74,8 @@
"recordLogName": "s3-recordlog"
},
"mongodb": {
"replicaSetHosts": "localhost:27017,localhost:27018,localhost:27019",
"writeConcern": "majority",
"replicaSet": "rs0",
"readPreference": "primary",
"database": "metadata"
},
"certFilePaths": {
"key": "tests/unit/testConfigs/allOptsConfig/key.txt",
"cert": "tests/unit/testConfigs/allOptsConfig/cert.txt",
"ca": "tests/unit/testConfigs/allOptsConfig/caBundle.txt"
},
"outboundProxy": {
"url": "http://test:8001",
"caBundle": "tests/unit/testConfigs/allOptsConfig/caBundle.txt",
"key": "tests/unit/testConfigs/allOptsConfig/key.txt",
"cert": "tests/unit/testConfigs/allOptsConfig/cert.txt"
},
"localCache": {
"name": "zenko",
"sentinels": "localhost:6379"
},
"redis": {
"name": "zenko",
"sentinels": "localhost:6379"
},
"scuba": {
"host": "localhost",
"port": 8100
},
"quota": {
"maxStalenessMS": 86400000
},
"utapi": {
"redis": {
"host": "localhost",
"port": 6379,
"retry": {
"connectBackoff": {
"min": 10,
"max": 1000,
"factor": 1.5,
"jitter": 0.1,
"deadline": 10000
}
}
}
},
"overlayVersion": 4
"host": "localhost",
"port": 27018,
"database": "metadata"
}
}

View File

@ -1,112 +1,82 @@
{
"us-east-1": {
"type": "file",
"objectId": "us-east-1",
"legacyAwsBehavior": true,
"details": {}
},
"us-east-2": {
"type": "file",
"objectId": "us-east-2",
"legacyAwsBehavior": false,
"details": {}
},
"us-west-1": {
"type": "file",
"objectId": "us-west-1",
"legacyAwsBehavior": false,
"details": {}
},
"us-west-2": {
"type": "file",
"objectId": "us-west-2",
"legacyAwsBehavior": false,
"details": {}
},
"ca-central-1": {
"type": "file",
"objectId": "ca-central-1",
"legacyAwsBehavior": false,
"details": {}
},
"cn-north-1": {
"type": "file",
"objectId": "cn-north-1",
"legacyAwsBehavior": false,
"details": {}
},
"ap-south-1": {
"type": "file",
"objectId": "ap-south-1",
"legacyAwsBehavior": false,
"details": {}
},
"ap-northeast-1": {
"type": "file",
"objectId": "ap-northeast-1",
"legacyAwsBehavior": false,
"details": {}
},
"ap-northeast-2": {
"type": "file",
"objectId": "ap-northeast-2",
"legacyAwsBehavior": false,
"details": {}
},
"ap-southeast-1": {
"type": "file",
"objectId": "ap-southeast-1",
"legacyAwsBehavior": false,
"details": {}
},
"ap-southeast-2": {
"type": "file",
"objectId": "ap-southeast-2",
"legacyAwsBehavior": false,
"details": {}
},
"eu-central-1": {
"type": "file",
"objectId": "eu-central-1",
"legacyAwsBehavior": false,
"details": {}
},
"eu-west-1": {
"type": "file",
"objectId": "eu-west-1",
"legacyAwsBehavior": false,
"details": {}
},
"eu-west-2": {
"type": "file",
"objectId": "eu-west-2",
"legacyAwsBehavior": false,
"details": {}
},
"EU": {
"type": "file",
"objectId": "EU",
"legacyAwsBehavior": false,
"details": {}
},
"sa-east-1": {
"type": "file",
"objectId": "sa-east-1",
"legacyAwsBehavior": false,
"details": {}
},
"location-dmf-v1": {
"type": "dmf",
"objectId": "location-dmf-v1",
"legacyAwsBehavior": false,
"isCold": true,
"details": {}
},
"location-azure-archive-v1": {
"type": "azure_archive",
"objectId": "location-azure-archive-v1",
"legacyAwsBehavior": false,
"isCold": true,
"details": {}
}
}

View File

@ -39,8 +39,6 @@ const constants = {
// once the multipart upload is complete.
mpuBucketPrefix: 'mpuShadowBucket',
blacklistedPrefixes: { bucket: [], object: [] },
// GCP Object Tagging Prefix
gcpTaggingPrefix: 'aws-tag-',
// PublicId is used as the canonicalID for a request that contains
// no authentication information. Requestor can access
// only public resources
@ -66,71 +64,48 @@ const constants = {
// http://docs.aws.amazon.com/AmazonS3/latest/API/mpUploadComplete.html
minimumAllowedPartSize: 5242880,
// AWS sets a maximum total parts limit
// https://docs.aws.amazon.com/AmazonS3/latest/API/mpUploadUploadPart.html
maximumAllowedPartCount: 10000,
gcpMaximumAllowedPartCount: 1024,
// Max size on put part or copy part is 5GB. For functional
// testing use 110 MB as max
maximumAllowedPartSize: process.env.MPU_TESTING === 'yes' ? 110100480 :
5368709120,
// Max size allowed in a single put object request is 5GB
// https://docs.aws.amazon.com/AmazonS3/latest/dev/UploadingObjects.html
maximumAllowedUploadSize: 5368709120,
// AWS states max size for user-defined metadata (x-amz-meta- headers) is
// 2 KB: http://docs.aws.amazon.com/AmazonS3/latest/API/RESTObjectPUT.html
// In testing, AWS seems to allow up to 88 more bytes, so we do the same.
maximumMetaHeadersSize: 2136,
// Maximum HTTP headers size allowed
maxHttpHeadersSize: 14122,
// hex digest of sha256 hash of empty string:
emptyStringHash: crypto.createHash('sha256')
.update('', 'binary').digest('hex'),
// Queries supported by AWS that we do not currently support.
// Non-bucket queries
unsupportedQueries: [
'accelerate',
'analytics',
'inventory',
'logging',
'metrics',
'policyStatus',
'publicAccessBlock',
'notification',
'policy',
'requestPayment',
'restore',
'torrent',
],
// Headers supported by AWS that we do not currently support.
unsupportedHeaders: [
'x-amz-server-side-encryption',
'x-amz-server-side-encryption-customer-algorithm',
'x-amz-server-side-encryption-aws-kms-key-id',
'x-amz-server-side-encryption-context',
'x-amz-server-side-encryption-customer-key',
'x-amz-server-side-encryption-customer-key-md5',
],
// user metadata header to set object locationConstraint
objectLocationConstraintHeader: 'x-amz-storage-class',
lastModifiedHeader: 'x-amz-meta-x-scal-last-modified',
objectLocationConstraintHeader: 'x-amz-meta-scal-location-constraint',
legacyLocations: ['sproxyd', 'legacy'],
// declare here all existing service accounts and their properties
// (if any, otherwise an empty object)
serviceAccountProperties: {
replication: {},
lifecycle: {},
gc: {},
'md-ingestion': {
canReplicate: true,
},
},
/* eslint-disable camelcase */
externalBackends: { aws_s3: true, azure: true, gcp: true, pfs: true, dmf: true, azure_archive: true },
externalBackends: { aws_s3: true, azure: true, gcp: true },
// some of the available data backends (if called directly rather
// than through the multiple backend gateway) need a key provided
// as a string as first parameter of the get/delete methods.
@ -139,110 +114,15 @@ const constants = {
// for external backends, don't call unless at least 1 minute
// (60,000 milliseconds) since last call
externalBackendHealthCheckInterval: 60000,
versioningNotImplBackends: { azure: true, gcp: true },
mpuMDStoredExternallyBackend: { aws_s3: true, gcp: true },
skipBatchDeleteBackends: { azure: true, gcp: true },
s3HandledBackends: { azure: true, gcp: true },
hasCopyPartBackends: { aws_s3: true, gcp: true },
versioningNotImplBackends: { azure: true },
mpuMDStoredExternallyBackend: { aws_s3: true },
/* eslint-enable camelcase */
mpuMDStoredOnS3Backend: { azure: true },
azureAccountNameRegex: /^[a-z0-9]{3,24}$/,
base64Regex: new RegExp('^(?:[A-Za-z0-9+/]{4})*' +
'(?:[A-Za-z0-9+/]{2}==|[A-Za-z0-9+/]{3}=)?$'),
productName: 'APN/1.0 Scality/1.0 Scality CloudServer for Zenko',
// location constraint delimiter
zenkoSeparator: ':',
// user metadata applied on zenko objects
zenkoIDHeader: 'x-amz-meta-zenko-instance-id',
bucketOwnerActions: [
'bucketDeleteCors',
'bucketDeleteLifecycle',
'bucketDeletePolicy',
'bucketDeleteReplication',
'bucketDeleteWebsite',
'bucketGetCors',
'bucketGetLifecycle',
'bucketGetLocation',
'bucketGetPolicy',
'bucketGetReplication',
'bucketGetVersioning',
'bucketGetWebsite',
'bucketPutCors',
'bucketPutLifecycle',
'bucketPutPolicy',
'bucketPutReplication',
'bucketPutVersioning',
'bucketPutWebsite',
'objectDeleteTagging',
'objectGetTagging',
'objectPutTagging',
'objectPutLegalHold',
'objectPutRetention',
],
// response header to be sent when there are invalid
// user metadata in the object's metadata
invalidObjectUserMetadataHeader: 'x-amz-missing-meta',
// Bucket specific queries supported by AWS that we do not currently support
// these queries may or may not be supported at object level
unsupportedBucketQueries: [
],
suppressedUtapiEventFields: [
'object',
'location',
'versionId',
],
allowedUtapiEventFilterFields: [
'operationId',
'location',
'account',
'user',
'bucket',
],
arrayOfAllowed: [
'objectPutTagging',
'objectPutLegalHold',
'objectPutRetention',
],
allowedUtapiEventFilterStates: ['allow', 'deny'],
allowedRestoreObjectRequestTierValues: ['Standard'],
lifecycleListing: {
CURRENT_TYPE: 'current',
NON_CURRENT_TYPE: 'noncurrent',
ORPHAN_DM_TYPE: 'orphan',
},
multiObjectDeleteConcurrency: 50,
maxScannedLifecycleListingEntries: 10000,
overheadField: [
'content-length',
'owner-id',
'versionId',
'isNull',
'isDeleteMarker',
],
unsupportedSignatureChecksums: new Set([
'STREAMING-UNSIGNED-PAYLOAD-TRAILER',
'STREAMING-AWS4-HMAC-SHA256-PAYLOAD-TRAILER',
'STREAMING-AWS4-ECDSA-P256-SHA256-PAYLOAD',
'STREAMING-AWS4-ECDSA-P256-SHA256-PAYLOAD-TRAILER',
]),
supportedSignatureChecksums: new Set([
'UNSIGNED-PAYLOAD',
'STREAMING-AWS4-HMAC-SHA256-PAYLOAD',
]),
ipv4Regex: /^(\d{1,3}\.){3}\d{1,3}(\/(3[0-2]|[12]?\d))?$/,
ipv6Regex: /^([\da-f]{1,4}:){7}[\da-f]{1,4}$/i,
// The AWS assumed Role resource type
assumedRoleArnResourceType: 'assumed-role',
// Session name of the backbeat lifecycle assumed role session.
backbeatLifecycleSessionName: 'backbeat-lifecycle',
actionsToConsiderAsObjectPut: [
'initiateMultipartUpload',
'objectPutPart',
'completeMultipartUpload',
],
// if requester is not bucket owner, bucket policy actions should be denied with
// MethodNotAllowed error
onlyOwnerAllowed: ['bucketDeletePolicy', 'bucketGetPolicy', 'bucketPutPolicy'],
};
module.exports = constants;

View File

@ -79,7 +79,7 @@ function createEncryptedBucket() {
.option('-b, --bucket <bucket>', 'Name of the bucket')
.option('-h, --host <host>', 'Host of the server')
.option('-p, --port <port>', 'Port of the server')
.option('-s, --ssl', 'Enable ssl')
.option('-s', '--ssl', 'Enable ssl')
.option('-v, --verbose')
.option('-l, --location-constraint <locationConstraint>',
'location Constraint')

View File

@ -8,6 +8,19 @@ const inMemory = require('./in_memory/backend').backend;
const file = require('./file/backend');
const KMIPClient = require('arsenal').network.kmipClient;
const Common = require('./common');
let scalityKMS;
let scalityKMSImpl;
try {
// eslint-disable-next-line import/no-unresolved
const ScalityKMS = require('scality-kms');
scalityKMS = new ScalityKMS(config.kms);
scalityKMSImpl = 'scalityKms';
} catch (err) {
logger.warn('scality kms unavailable. ' +
'Using file kms backend unless mem specified.');
scalityKMS = file;
scalityKMSImpl = 'fileKms';
}
let client;
let implName;
@ -19,9 +32,8 @@ if (config.backends.kms === 'mem') {
client = file;
implName = 'fileKms';
} else if (config.backends.kms === 'scality') {
const ScalityKMS = require('scality-kms');
client = new ScalityKMS(config.kms);
implName = 'scalityKms';
client = scalityKMS;
implName = scalityKMSImpl;
} else if (config.backends.kms === 'kmip') {
const kmipConfig = { kmip: config.kmip };
if (!kmipConfig.kmip) {
@ -57,45 +69,52 @@ class KMS {
/**
*
* @param {string} bucketName - bucket name
* @param {object} sseConfig - SSE configuration
* @param {object} headers - request headers
* @param {object} log - logger object
* @param {function} cb - callback
* @returns {undefined}
* @callback called with (err, serverSideEncryptionInfo: object)
*/
static bucketLevelEncryption(bucketName, sseConfig, log, cb) {
static bucketLevelEncryption(bucketName, headers, log, cb) {
const sseAlgorithm = headers['x-amz-scal-server-side-encryption'];
const sseMasterKeyId =
headers['x-amz-scal-server-side-encryption-aws-kms-key-id'];
/*
The purpose of bucket level encryption is so that the client does not
have to send appropriate headers to trigger encryption on each object
put in an "encrypted bucket". Customer provided keys are not
feasible in this system because we do not want to store this key
in the bucket metadata.
*/
const { algorithm, configuredMasterKeyId, mandatory } = sseConfig;
const _mandatory = mandatory === true;
if (algorithm === 'AES256' || algorithm === 'aws:kms') {
return this.createBucketKey(bucketName, log, (err, masterKeyId) => {
*/
if (sseAlgorithm === 'AES256' ||
(sseAlgorithm === 'aws:kms' && sseMasterKeyId === undefined)) {
this.createBucketKey(bucketName, log, (err, masterKeyId) => {
if (err) {
return cb(err);
cb(err);
return;
}
const serverSideEncryptionInfo = {
cryptoScheme: 1,
algorithm,
algorithm: sseAlgorithm,
masterKeyId,
mandatory: _mandatory,
mandatory: true,
};
if (algorithm === 'aws:kms' && configuredMasterKeyId) {
serverSideEncryptionInfo.configuredMasterKeyId = configuredMasterKeyId;
}
return cb(null, serverSideEncryptionInfo);
cb(null, serverSideEncryptionInfo);
});
} else if (sseAlgorithm === 'aws:kms') {
const serverSideEncryptionInfo = {
cryptoScheme: 1,
algorithm: sseAlgorithm,
masterKeyId: sseMasterKeyId,
mandatory: true,
};
cb(null, serverSideEncryptionInfo);
} else {
/*
* no encryption
*/
cb(null, null);
}
/*
* no encryption
*/
return cb(null, null);
}
/**
@ -150,18 +169,9 @@ class KMS {
static createCipherBundle(serverSideEncryptionInfo,
log, cb) {
const dataKey = this.createDataKey(log);
const { algorithm, configuredMasterKeyId, masterKeyId: bucketMasterKeyId } = serverSideEncryptionInfo;
let masterKeyId = bucketMasterKeyId;
if (configuredMasterKeyId) {
log.debug('using user configured kms master key id');
masterKeyId = configuredMasterKeyId;
}
const cipherBundle = {
algorithm,
masterKeyId,
algorithm: serverSideEncryptionInfo.algorithm,
masterKeyId: serverSideEncryptionInfo.masterKeyId,
cryptoScheme: 1,
cipheredDataKey: null,
cipher: null,
@ -171,7 +181,7 @@ class KMS {
function cipherDataKey(next) {
log.debug('ciphering a data key');
return client.cipherDataKey(cipherBundle.cryptoScheme,
cipherBundle.masterKeyId,
serverSideEncryptionInfo.masterKeyId,
dataKey, log, (err, cipheredDataKey) => {
if (err) {
log.debug('error from kms',
@ -286,29 +296,6 @@ class KMS {
return cb(err, decipherBundle);
});
}
static checkHealth(log, cb) {
if (!client.healthcheck) {
return cb(null, {
[implName]: { code: 200, message: 'OK' },
});
}
return client.healthcheck(log, err => {
const respBody = {};
if (err) {
respBody[implName] = {
error: err.description,
code: err.code,
};
} else {
respBody[implName] = {
code: 200,
message: 'OK',
};
}
return cb(null, respBody);
});
}
}
module.exports = KMS;

18
core/utapi/utapi.js Normal file
View File

@ -0,0 +1,18 @@
const utapiServer = require('utapi').UtapiServer;
const _config = require('../Config').config;
// start utapi server
if (_config.utapi) {
const fullConfig = Object.assign({}, _config.utapi);
if (_config.vaultd) {
Object.assign(fullConfig, { vaultd: _config.vaultd });
}
if (_config.https) {
Object.assign(fullConfig, { https: _config.https });
}
// copy healthcheck IPs
if (_config.healthChecks) {
Object.assign(fullConfig, { healthChecks: _config.healthChecks });
}
utapiServer(fullConfig);
}

View File

@ -0,0 +1,6 @@
const UtapiReplay = require('utapi').UtapiReplay;
const _config = require('../Config').config;
// start utapi server
const replay = new UtapiReplay(_config.utapi);
replay.start();

View File

@ -2,72 +2,23 @@ const http = require('http');
const https = require('https');
const commander = require('commander');
const { auth } = require('arsenal');
const { UtapiClient, utapiVersion } = require('utapi');
const { UtapiClient } = require('utapi');
const logger = require('../utilities/logger');
const _config = require('../Config').config;
const { suppressedUtapiEventFields: suppressedEventFields } = require('../../constants');
// setup utapi client
let utapiConfig;
if (utapiVersion === 1 && _config.utapi) {
if (_config.utapi.redis === undefined) {
utapiConfig = Object.assign({}, _config.utapi, { redis: _config.redis });
} else {
utapiConfig = Object.assign({}, _config.utapi);
}
} else if (utapiVersion === 2) {
utapiConfig = Object.assign({
tls: _config.https,
suppressedEventFields,
}, _config.utapi || {});
}
const utapi = new UtapiClient(utapiConfig);
const bucketOwnerMetrics = [
'completeMultipartUpload',
'multiObjectDelete',
'abortMultipartUpload',
'copyObject',
'deleteObject',
'putObject',
'uploadPartCopy',
'uploadPart',
];
function evalAuthInfo(authInfo, canonicalID, action) {
let accountId = authInfo.getCanonicalID();
let userId = authInfo.isRequesterAnIAMUser() ?
authInfo.getShortid() : undefined;
// If action impacts 'numberOfObjectsStored' or 'storageUtilized' metric
// only the bucket owner account's metrics should be updated
const canonicalIdMatch = authInfo.getCanonicalID() === canonicalID;
if (bucketOwnerMetrics.includes(action) && !canonicalIdMatch) {
accountId = canonicalID;
userId = undefined;
}
return {
accountId,
userId,
};
}
const utapi = new UtapiClient(_config.utapi);
function _listMetrics(host,
port,
metric,
metricType,
timeRange,
accessKey,
secretKey,
verbose,
recent,
ssl) {
port,
metric,
metricType,
timeRange,
accessKey,
secretKey,
verbose,
recent,
ssl) {
const listAction = recent ? 'ListRecentMetrics' : 'ListMetrics';
// If recent listing, we do not provide `timeRange` in the request
const requestObj = recent
? { [metric]: metricType }
: { timeRange, [metric]: metricType };
const requestBody = JSON.stringify(requestObj);
const options = {
host,
port,
@ -76,7 +27,6 @@ function _listMetrics(host,
headers: {
'content-type': 'application/json',
'cache-control': 'no-cache',
'content-length': Buffer.byteLength(requestBody),
},
rejectUnauthorized: false,
};
@ -114,8 +64,10 @@ function _listMetrics(host,
if (verbose) {
logger.info('request headers', { headers: request._headers });
}
request.write(requestBody);
// If recent listing, we do not provide `timeRange` in the request
const requestObj = recent ? {} : { timeRange };
requestObj[metric] = metricType;
request.write(JSON.stringify(requestObj));
request.end();
}
@ -138,7 +90,7 @@ function listMetrics(metricType) {
if (metricType === 'buckets') {
commander
.option('-b, --buckets <buckets>', 'Name of bucket(s) with ' +
'a comma separator if more than one');
'a comma separator if more than one');
} else {
commander
.option('-m, --metric <metric>', 'Metric type')
@ -193,19 +145,6 @@ function listMetrics(metricType) {
}
});
// The string `commander[metric]` is a comma-separated list of resources
// given by the user.
const resources = commander[metric].split(',');
// Validate passed accounts to remove any canonicalIDs
if (metric === 'accounts') {
const invalid = resources.filter(r => !/^\d{12}$/.test(r));
if (invalid.length > 0) {
logger.error(`Invalid account ID: ${invalid.join(', ')}`);
process.exit(1);
}
}
const timeRange = [];
// If recent listing, we disregard any start or end option given
if (!recent) {
@ -228,7 +167,9 @@ function listMetrics(metricType) {
timeRange.push(numEnd);
}
}
// The string `commander[metric]` is a comma-separated list of resources
// given by the user.
const resources = commander[metric].split(',');
_listMetrics(host, port, metric, resources, timeRange, accessKey, secretKey,
verbose, recent, ssl);
}
@ -249,84 +190,13 @@ function listMetrics(metricType) {
* @param {number} [metricObj.newByteLength] - (optional) new object size
* @param {number|null} [metricObj.oldByteLength] - (optional) old object size
* (obj. overwrites)
* @param {number} [metricObj.numberOfObjects] - (optional) number of objects
* @param {number} [metricObj.numberOfObjects] - (optional) number of obects
* added/deleted
* @param {boolean} [metricObject.isDelete] - (optional) Indicates whether this
* is a delete operation
* @return {function | undefined} - `utapi.pushMetric` or undefined if the action is
* filtered out and not pushed to utapi.
* @return {function} - `utapi.pushMetric`
*/
function pushMetric(action, log, metricObj) {
const {
bucket,
keys,
versionId,
byteLength,
newByteLength,
oldByteLength,
numberOfObjects,
authInfo,
canonicalID,
location,
isDelete,
removedDeleteMarkers,
} = metricObj;
if (utapiVersion === 2) {
const incomingBytes = action === 'getObject' ? 0 : newByteLength;
let sizeDelta = incomingBytes;
if (Number.isInteger(oldByteLength) && Number.isInteger(newByteLength)) {
sizeDelta = newByteLength - oldByteLength;
// Include oldByteLength in conditional so we don't end up with `-0`
} else if (action === 'completeMultipartUpload' && !versionId && oldByteLength) {
// If this is a non-versioned bucket we need to decrement
// the sizeDelta added by uploadPart when completeMPU is called.
sizeDelta = -oldByteLength;
} else if (action === 'abortMultipartUpload' && byteLength) {
sizeDelta = -byteLength;
} else if (action === 'putDeleteMarkerObject' && byteLength) {
sizeDelta = -byteLength;
}
let objectDelta = isDelete ? -numberOfObjects : numberOfObjects;
// putDeleteMarkerObject does not pass numberOfObjects
if ((action === 'putDeleteMarkerObject' && byteLength === null)
|| action === 'replicateDelete'
|| action === 'replicateObject') {
objectDelta = 1;
} else if (action === 'multiObjectDelete') {
objectDelta = -(numberOfObjects + removedDeleteMarkers);
}
const utapiObj = {
operationId: action,
bucket,
location,
objectDelta,
sizeDelta: isDelete ? -byteLength : sizeDelta,
incomingBytes,
outgoingBytes: action === 'getObject' ? newByteLength : 0,
};
// Any operation from lifecycle that does not change object count or size is dropped
const isLifecycle = _config.lifecycleRoleName
&& authInfo && authInfo.arn.endsWith(`:assumed-role/${_config.lifecycleRoleName}/backbeat-lifecycle`);
if (isLifecycle && !objectDelta && !sizeDelta) {
log.trace('ignoring pushMetric from lifecycle service user', { action, bucket, keys });
return undefined;
}
if (keys && keys.length === 1) {
[utapiObj.object] = keys;
if (versionId) {
utapiObj.versionId = versionId;
}
}
utapiObj.account = authInfo ? evalAuthInfo(authInfo, canonicalID, action).accountId : canonicalID;
utapiObj.user = authInfo ? evalAuthInfo(authInfo, canonicalID, action).userId : undefined;
return utapi.pushMetric(utapiObj);
}
const { bucket, keys, byteLength, newByteLength,
oldByteLength, numberOfObjects, authInfo, canonicalID } = metricObj;
const utapiObj = {
bucket,
keys,
@ -339,56 +209,16 @@ function pushMetric(action, log, metricObj) {
// account-level metrics and the shortId for user-level metrics. Otherwise
// check if the canonical ID is already provided for account-level metrics.
if (authInfo) {
const { accountId, userId } = evalAuthInfo(authInfo, canonicalID, action);
utapiObj.accountId = accountId;
utapiObj.userId = userId;
utapiObj.accountId = authInfo.getCanonicalID();
utapiObj.userId = authInfo.isRequesterAnIAMUser() ?
authInfo.getShortid() : undefined;
} else if (canonicalID) {
utapiObj.accountId = canonicalID;
}
return utapi.pushMetric(action, log.getSerializedUids(), utapiObj);
}
/**
* internal: get the unique location ID from the location name
*
* @param {string} location - location name
* @return {string} - location unique ID
*/
function _getLocationId(location) {
return _config.locationConstraints[location].objectId;
}
/**
* Call the Utapi Client 'getLocationMetric' method with the
* associated parameters
* @param {string} location - name of data backend to list metric for
* @param {object} log - werelogs logger
* @param {function} cb - callback to call
* @return {function} - `utapi.getLocationMetric`
*/
function getLocationMetric(location, log, cb) {
const locationId = _getLocationId(location);
return utapi.getLocationMetric(locationId, log.getSerializedUids(), cb);
}
/**
* Call the Utapi Client 'pushLocationMetric' method with the
* associated parameters
* @param {string} location - name of data backend
* @param {number} byteLength - number of bytes
* @param {object} log - werelogs logger
* @param {function} cb - callback to call
* @return {function} - `utapi.pushLocationMetric`
*/
function pushLocationMetric(location, byteLength, log, cb) {
const locationId = _getLocationId(location);
return utapi.pushLocationMetric(locationId, byteLength,
log.getSerializedUids(), cb);
}
module.exports = {
listMetrics,
pushMetric,
getLocationMetric,
pushLocationMetric,
};

View File

@ -1,220 +0,0 @@
#!/bin/bash
# set -e stops the execution of a script if a command or pipeline has an error
set -e
# modifying config.json
JQ_FILTERS_CONFIG="."
# ENDPOINT var can accept comma separated values
# for multiple endpoint locations
if [[ "$ENDPOINT" ]]; then
IFS="," read -ra HOST_NAMES <<< "$ENDPOINT"
for host in "${HOST_NAMES[@]}"; do
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .restEndpoints[\"$host\"]=\"us-east-1\""
done
echo "Host name has been modified to ${HOST_NAMES[@]}"
echo "Note: In your /etc/hosts file on Linux, OS X, or Unix with root permissions, make sure to associate 127.0.0.1 with ${HOST_NAMES[@]}"
fi
if [[ "$LOG_LEVEL" ]]; then
if [[ "$LOG_LEVEL" == "info" || "$LOG_LEVEL" == "debug" || "$LOG_LEVEL" == "trace" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .log.logLevel=\"$LOG_LEVEL\""
echo "Log level has been modified to $LOG_LEVEL"
else
echo "The log level you provided is incorrect (info/debug/trace)"
fi
fi
if [[ "$SSL" && "$HOST_NAMES" ]]; then
# This condition makes sure that the certificates are not generated twice. (for docker restart)
if [ ! -f ./ca.key ] || [ ! -f ./ca.crt ] || [ ! -f ./server.key ] || [ ! -f ./server.crt ] ; then
# Compute config for utapi tests
cat >>req.cfg <<EOF
[req]
distinguished_name = req_distinguished_name
prompt = no
req_extensions = s3_req
[req_distinguished_name]
CN = ${HOST_NAMES[0]}
[s3_req]
subjectAltName = @alt_names
extendedKeyUsage = serverAuth, clientAuth
[alt_names]
DNS.1 = *.${HOST_NAMES[0]}
DNS.2 = ${HOST_NAMES[0]}
EOF
## Generate SSL key and certificates
# Generate a private key for your CSR
openssl genrsa -out ca.key 2048
# Generate a self signed certificate for your local Certificate Authority
openssl req -new -x509 -extensions v3_ca -key ca.key -out ca.crt -days 99999 -subj "/C=US/ST=Country/L=City/O=Organization/CN=S3 CA Server"
# Generate a key for S3 Server
openssl genrsa -out server.key 2048
# Generate a Certificate Signing Request for S3 Server
openssl req -new -key server.key -out server.csr -config req.cfg
# Generate a local-CA-signed certificate for S3 Server
openssl x509 -req -in server.csr -CA ca.crt -CAkey ca.key -CAcreateserial -out server.crt -days 99999 -sha256 -extfile req.cfg -extensions s3_req
fi
## Update S3Server config.json
# This condition makes sure that certFilePaths section is not added twice. (for docker restart)
if ! grep -q "certFilePaths" ./config.json; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .certFilePaths= { \"key\": \".\/server.key\", \"cert\": \".\/server.crt\", \"ca\": \".\/ca.crt\" }"
fi
fi
if [[ "$LISTEN_ADDR" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .metadataDaemon.bindAddress=\"$LISTEN_ADDR\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .dataDaemon.bindAddress=\"$LISTEN_ADDR\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .pfsDaemon.bindAddress=\"$LISTEN_ADDR\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .listenOn=[\"$LISTEN_ADDR:8000\"]"
fi
if [[ "$REPLICATION_GROUP_ID" ]] ; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .replicationGroupId=\"$REPLICATION_GROUP_ID\""
fi
if [[ "$DATA_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .dataClient.host=\"$DATA_HOST\""
fi
if [[ "$METADATA_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .metadataClient.host=\"$METADATA_HOST\""
fi
if [[ "$PFSD_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .pfsClient.host=\"$PFSD_HOST\""
fi
if [[ "$MONGODB_HOSTS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .mongodb.replicaSetHosts=\"$MONGODB_HOSTS\""
fi
if [[ "$MONGODB_RS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .mongodb.replicaSet=\"$MONGODB_RS\""
fi
if [[ "$MONGODB_DATABASE" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .mongodb.database=\"$MONGODB_DATABASE\""
fi
if [ -z "$REDIS_HA_NAME" ]; then
REDIS_HA_NAME='mymaster'
fi
if [[ "$REDIS_SENTINELS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.name=\"$REDIS_HA_NAME\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.sentinels=\"$REDIS_SENTINELS\""
elif [[ "$REDIS_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.host=\"$REDIS_HOST\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.port=6379"
fi
if [[ "$REDIS_PORT" ]] && [[ ! "$REDIS_SENTINELS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.port=$REDIS_PORT"
fi
if [[ "$REDIS_SENTINELS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .redis.name=\"$REDIS_HA_NAME\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .redis.sentinels=\"$REDIS_SENTINELS\""
elif [[ "$REDIS_HA_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .redis.host=\"$REDIS_HA_HOST\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .redis.port=6379"
fi
if [[ "$REDIS_HA_PORT" ]] && [[ ! "$REDIS_SENTINELS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .redis.port=$REDIS_HA_PORT"
fi
if [[ "$RECORDLOG_ENABLED" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .recordLog.enabled=true"
fi
if [[ "$STORAGE_LIMIT_ENABLED" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .utapi.metrics[.utapi.metrics | length]=\"location\""
fi
if [[ "$CRR_METRICS_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .backbeat.host=\"$CRR_METRICS_HOST\""
fi
if [[ "$CRR_METRICS_PORT" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .backbeat.port=$CRR_METRICS_PORT"
fi
if [[ "$WE_OPERATOR_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .workflowEngineOperator.host=\"$WE_OPERATOR_HOST\""
fi
if [[ "$WE_OPERATOR_PORT" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .workflowEngineOperator.port=$WE_OPERATOR_PORT"
fi
if [[ "$HEALTHCHECKS_ALLOWFROM" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .healthChecks.allowFrom=[\"$HEALTHCHECKS_ALLOWFROM\"]"
fi
# external backends http(s) agent config
# AWS
if [[ "$AWS_S3_HTTPAGENT_KEEPALIVE" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.aws_s3.httpAgent.keepAlive=$AWS_S3_HTTPAGENT_KEEPALIVE"
fi
if [[ "$AWS_S3_HTTPAGENT_KEEPALIVE_MS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.aws_s3.httpAgent.keepAliveMsecs=$AWS_S3_HTTPAGENT_KEEPALIVE_MS"
fi
if [[ "$AWS_S3_HTTPAGENT_KEEPALIVE_MAX_SOCKETS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.aws_s3.httpAgent.maxSockets=$AWS_S3_HTTPAGENT_KEEPALIVE_MAX_SOCKETS"
fi
if [[ "$AWS_S3_HTTPAGENT_KEEPALIVE_MAX_FREE_SOCKETS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.aws_s3.httpAgent.maxFreeSockets=$AWS_S3_HTTPAGENT_KEEPALIVE_MAX_FREE_SOCKETS"
fi
#GCP
if [[ "$GCP_HTTPAGENT_KEEPALIVE" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.gcp.httpAgent.keepAlive=$GCP_HTTPAGENT_KEEPALIVE"
fi
if [[ "$GCP_HTTPAGENT_KEEPALIVE_MS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.gcp.httpAgent.keepAliveMsecs=$GCP_HTTPAGENT_KEEPALIVE_MS"
fi
if [[ "$GCP_HTTPAGENT_KEEPALIVE_MAX_SOCKETS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.gcp.httpAgent.maxSockets=$GCP_HTTPAGENT_KEEPALIVE_MAX_SOCKETS"
fi
if [[ "$GCP_HTTPAGENT_KEEPALIVE_MAX_FREE_SOCKETS" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .externalBackends.gcp.httpAgent.maxFreeSockets=$GCP_HTTPAGENT_KEEPALIVE_MAX_FREE_SOCKETS"
fi
if [[ -n "$BUCKET_DENY_FILTER" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .utapi.filter.deny.bucket=[\"$BUCKET_DENY_FILTER\"]"
fi
if [[ "$TESTING_MODE" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .testingMode=true"
fi
if [[ $JQ_FILTERS_CONFIG != "." ]]; then
jq "$JQ_FILTERS_CONFIG" config.json > config.json.tmp
mv config.json.tmp config.json
fi
if test -v INITIAL_INSTANCE_ID && test -v S3METADATAPATH && ! test -f ${S3METADATAPATH}/uuid ; then
echo -n ${INITIAL_INSTANCE_ID} > ${S3METADATAPATH}/uuid
fi
# s3 secret credentials for Zenko
if [ -r /run/secrets/s3-credentials ] ; then
. /run/secrets/s3-credentials
fi
exec "$@"

25
docker/Dockerfile Normal file
View File

@ -0,0 +1,25 @@
FROM node:6-slim
MAINTAINER Giorgio Regni <gr@scality.com>
WORKDIR /usr/src/app
# Keep the .git directory in order to properly report version
COPY ./package.json .
RUN apt-get update \
&& apt-get install -y jq python git build-essential --no-install-recommends \
&& npm install --production \
&& apt-get autoremove --purge -y python git build-essential \
&& rm -rf /var/lib/apt/lists/* \
&& npm cache clear \
&& rm -rf ~/.node-gyp \
&& rm -rf /tmp/npm-*
COPY ./ ./
VOLUME ["/usr/src/app/localData","/usr/src/app/localMetadata"]
ENTRYPOINT ["/usr/src/app/docker-entrypoint.sh"]
CMD [ "npm", "start" ]
EXPOSE 8000

View File

@ -7,16 +7,16 @@ COPY . /usr/src/app
RUN apt-get update \
&& apt-get install -y jq python git build-essential --no-install-recommends \
&& yarn install --production \
&& npm install --production \
&& apt-get autoremove --purge -y python git build-essential \
&& rm -rf /var/lib/apt/lists/* \
&& yarn cache clean \
&& npm cache clear \
&& rm -rf ~/.node-gyp \
&& rm -rf /tmp/yarn-*
&& rm -rf /tmp/npm-*
ENV S3BACKEND mem
ENTRYPOINT ["/usr/src/app/docker-entrypoint.sh"]
CMD [ "yarn", "start" ]
CMD [ "npm", "start" ]
EXPOSE 8000

107
docker/docker-entrypoint.sh Executable file
View File

@ -0,0 +1,107 @@
#!/bin/bash
# set -e stops the execution of a script if a command or pipeline has an error
set -e
# modifying config.json
JQ_FILTERS_CONFIG="."
if [[ "$ENDPOINT" ]]; then
HOST_NAME="$ENDPOINT"
fi
if [[ "$HOST_NAME" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .restEndpoints[\"$HOST_NAME\"]=\"us-east-1\""
echo "Host name has been modified to $HOST_NAME"
echo "Note: In your /etc/hosts file on Linux, OS X, or Unix with root permissions, make sure to associate 127.0.0.1 with $HOST_NAME"
fi
if [[ "$LOG_LEVEL" ]]; then
if [[ "$LOG_LEVEL" == "info" || "$LOG_LEVEL" == "debug" || "$LOG_LEVEL" == "trace" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .log.logLevel=\"$LOG_LEVEL\""
echo "Log level has been modified to $LOG_LEVEL"
else
echo "The log level you provided is incorrect (info/debug/trace)"
fi
fi
if [[ "$SSL" && "$HOST_NAME" ]]; then
# This condition makes sure that the certificates are not generated twice. (for docker restart)
if [ ! -f ./ca.key ] || [ ! -f ./ca.crt ] || [ ! -f ./server.key ] || [ ! -f ./server.crt ] ; then
# Compute config for utapi tests
cat >>req.cfg <<EOF
[req]
distinguished_name = req_distinguished_name
prompt = no
req_extensions = s3_req
[req_distinguished_name]
CN = ${HOST_NAME}
[s3_req]
subjectAltName = @alt_names
extendedKeyUsage = serverAuth, clientAuth
[alt_names]
DNS.1 = *.${HOST_NAME}
DNS.2 = ${HOST_NAME}
EOF
## Generate SSL key and certificates
# Generate a private key for your CSR
openssl genrsa -out ca.key 2048
# Generate a self signed certificate for your local Certificate Authority
openssl req -new -x509 -extensions v3_ca -key ca.key -out ca.crt -days 99999 -subj "/C=US/ST=Country/L=City/O=Organization/CN=S3 CA Server"
# Generate a key for S3 Server
openssl genrsa -out server.key 2048
# Generate a Certificate Signing Request for S3 Server
openssl req -new -key server.key -out server.csr -config req.cfg
# Generate a local-CA-signed certificate for S3 Server
openssl x509 -req -in server.csr -CA ca.crt -CAkey ca.key -CAcreateserial -out server.crt -days 99999 -sha256 -extfile req.cfg -extensions s3_req
fi
## Update S3Server config.json
# This condition makes sure that certFilePaths section is not added twice. (for docker restart)
if ! grep -q "certFilePaths" ./config.json; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .certFilePaths= { \"key\": \".\/server.key\", \"cert\": \".\/server.crt\", \"ca\": \".\/ca.crt\" }"
fi
fi
if [[ "$LISTEN_ADDR" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .metadataDaemon.bindAddress=\"$LISTEN_ADDR\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .dataDaemon.bindAddress=\"$LISTEN_ADDR\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .listenOn=[\"$LISTEN_ADDR:8000\"]"
fi
if [[ "$DATA_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .dataClient.host=\"$DATA_HOST\""
fi
if [[ "$METADATA_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .metadataClient.host=\"$METADATA_HOST\""
fi
if [[ "$REDIS_HOST" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.host=\"$REDIS_HOST\""
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.port=6379"
fi
if [[ "$REDIS_PORT" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .localCache.port=$REDIS_PORT"
fi
if [[ "$RECORDLOG_ENABLED" ]]; then
JQ_FILTERS_CONFIG="$JQ_FILTERS_CONFIG | .recordLog.enabled=true"
fi
if [[ $JQ_FILTERS_CONFIG != "." ]]; then
jq "$JQ_FILTERS_CONFIG" config.json > config.json.tmp
mv config.json.tmp config.json
fi
# s3 secret credentials for Zenko
if [ -r /run/secrets/s3-credentials ] ; then
. /run/secrets/s3-credentials
fi
exec "$@"

View File

@ -27,7 +27,7 @@ including null versions and delete markers, described in the above
links.
Implementation of Bucket Versioning in Zenko CloudServer
--------------------------------------------------------
-----------------------------------------
Overview of Metadata and API Component Roles
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
@ -66,7 +66,7 @@ The second section, `"Implementation of Bucket Versioning in
API" <#implementation-of-bucket-versioning-in-api>`__, describes the way
the metadata options are used in the API within S3 actions to create new
versions, update their metadata, and delete them. The management of null
versions and creation of delete markers is also described in this
versions and creation of delete markers are also described in this
section.
Implementation of Bucket Versioning in Metadata
@ -179,13 +179,12 @@ PUT
the master version with this version.
- ``versionId: <versionId>`` create or update a specific version (for updating
version's ACL or tags, or remote updates in geo-replication)
* if the version identified by ``versionId`` happens to be the latest
- if the version identified by ``versionId`` happens to be the latest
version, the master version will be updated as well
* if the master version is not as recent as the version identified by
- if the master version is not as recent as the version identified by
``versionId``, as may happen with cross-region replication, the master
will be updated as well
* note that with ``versionId`` set to an empty string ``''``, it will
- note that with ``versionId`` set to an empty string ``''``, it will
overwrite the master version only (same as no options, but the master
version will have a ``versionId`` property set in its metadata like
any other version). The ``versionId`` will never be exposed to an
@ -209,13 +208,10 @@ A deletion targeting the latest version of an object has to:
- delete the specified version identified by ``versionId``
- replace the master version with a version that is a placeholder for
deletion
- this version contains a special keyword, 'isPHD', to indicate the
master version was deleted and needs to be updated
- initiate a repair operation to update the value of the master
version:
- involves listing the versions of the object and get the latest
version to replace the placeholder delete version
- if no more versions exist, metadata deletes the master version,
@ -746,7 +742,7 @@ Operation
Startup
~~~~~~~
The simplest deployment is still to launch with yarn start, this will
The simplest deployment is still to launch with npm start, this will
start one instance of the Zenko CloudServer connector and will listen on the
locally bound dmd ports 9990 and 9991 (by default, see below).
@ -755,22 +751,22 @@ command in the Zenko CloudServer directory:
::
yarn run start_dmd
npm run start_dmd
This will open two ports:
- one is based on socket.io and is used for metadata transfers (9990 by
- one is based on socket.io and is used for metadata transfers (9990 by
default)
- the other is a REST interface used for data transfers (9991 by
- the other is a REST interface used for data transfers (9991 by
default)
Then, one or more instances of Zenko CloudServer without the dmd can be started
elsewhere with:
.. code:: sh
::
yarn run start_s3server
npm run start_s3server
Configuration
~~~~~~~~~~~~~
@ -796,10 +792,10 @@ access:
To run a remote dmd, you have to do the following:
- change both ``"host"`` attributes to the IP or host name where the
- change both ``"host"`` attributes to the IP or host name where the
dmd is run.
- Modify the ``"bindAddress"`` attributes in ``"metadataDaemon"`` and
- Modify the ``"bindAddress"`` attributes in ``"metadataDaemon"`` and
``"dataDaemon"`` sections where the dmd is run to accept remote
connections (e.g. ``"::"``)
@ -835,13 +831,13 @@ and ``createReadStream``. They more or less map the parameters accepted
by the corresponding calls in the LevelUp implementation of LevelDB.
They differ in the following:
- The ``sync`` option is ignored (under the hood, puts are gathered
- The ``sync`` option is ignored (under the hood, puts are gathered
into batches which have their ``sync`` property enforced when they
are committed to the storage)
- Some additional versioning-specific options are supported
- Some additional versioning-specific options are supported
- ``createReadStream`` becomes asynchronous, takes an additional
- ``createReadStream`` becomes asynchronous, takes an additional
callback argument and returns the stream in the second callback
parameter
@ -851,10 +847,10 @@ with ``DEBUG='socket.io*'`` environment variable set.
One parameter controls the timeout value after which RPC commands sent
end with a timeout error, it can be changed either:
- via the ``DEFAULT_CALL_TIMEOUT_MS`` option in
- via the ``DEFAULT_CALL_TIMEOUT_MS`` option in
``lib/network/rpc/rpc.js``
- or in the constructor call of the ``MetadataFileClient`` object (in
- or in the constructor call of the ``MetadataFileClient`` object (in
``lib/metadata/bucketfile/backend.js`` as ``callTimeoutMs``.
Default value is 30000.
@ -868,10 +864,10 @@ can tune the behavior (for better throughput or getting it more robust
on weak networks), they have to be set in ``mdserver.js`` file directly,
as there is no support in ``config.json`` for now for those options:
- ``streamMaxPendingAck``: max number of pending ack events not yet
- ``streamMaxPendingAck``: max number of pending ack events not yet
received (default is 5)
- ``streamAckTimeoutMs``: timeout for receiving an ack after an output
- ``streamAckTimeoutMs``: timeout for receiving an ack after an output
stream packet is sent to the client (default is 5000)
Data exchange through the REST data port
@ -922,17 +918,17 @@ Listing Types
We use three different types of metadata listing for various operations.
Here are the scenarios we use each for:
- 'Delimiter' - when no versions are possible in the bucket since it is
- 'Delimiter' - when no versions are possible in the bucket since it is
an internally-used only bucket which is not exposed to a user.
Namely,
1. to list objects in the "user's bucket" to respond to a GET SERVICE
request and
request and
2. to do internal listings on an MPU shadow bucket to complete multipart
upload operations.
upload operations.
- 'DelimiterVersion' - to list all versions in a bucket
- 'DelimiterMaster' - to list just the master versions of objects in a
- 'DelimiterVersion' - to list all versions in a bucket
- 'DelimiterMaster' - to list just the master versions of objects in a
bucket
Algorithms

View File

@ -1,146 +0,0 @@
# Bucket Policy Documentation
## Description
Bucket policy is a method of controlling access to a user's account at the
resource level.
There are three associated APIs:
- PUT Bucket policy (see https://docs.aws.amazon.com/AmazonS3/latest/API/RESTBucketPUTpolicy.html)
- GET Bucket policy (see https://docs.aws.amazon.com/AmazonS3/latest/API/RESTBucketGETpolicy.html)
- DELETE Bucket policy (see https://docs.aws.amazon.com/AmazonS3/latest/API/RESTBucketDELETEpolicy.html)
More information on bucket policies in general can be found at
https://docs.aws.amazon.com/AmazonS3/latest/dev/using-iam-policies.html.
## Requirements
To prevent loss of access to a bucket, the root owner of a bucket will always
be able to perform any of the three bucket policy-related operations, even
if permission is explicitly denied.
All other users must have permission to perform the desired operation.
## Design
On a PUTBucketPolicy request, the user provides a policy in JSON format.
The policy is evaluated against our policy schema in Arsenal and, once
validated, is stored as part of the bucket's metadata.
On a GETBucketPolicy request, the policy is retrieved from the bucket's
metadata.
On a DELETEBucketPolicy request, the policy is deleted from the bucket's
metadata.
All other APIs are updated to check if a bucket policy is attached to the bucket
the request is made on. If there is a policy, user authorization to perform
the requested action is checked.
### Differences Between Bucket and IAM Policies
IAM policies are attached to an IAM identity and define what actions that
identity is allowed to or denied from doing on what resource.
Bucket policies attach only to buckets and define what actions are allowed or
denied for which principles on that bucket. Permissions specified in a bucket
policy apply to all objects in that bucket unless otherwise specified.
Besides their attachment origins, the main structural difference between
IAM policy and bucket policy is the requirement of a "Principal" element in
bucket policies. This field is redundant in IAM policies.
### Policy Validation
For general guidelines for bucket policy structure, see examples here:
https://docs.aws.amazon.com/AmazonS3/latest/dev//example-bucket-policies.html.
Each bucket policy statement object requires at least four keys:
"Effect", "Principle", "Resource", and "Action".
"Effect" defines the effect of the policy and can have a string value of either
"Allow" or "Deny".
"Resource" defines to which bucket or list of buckets a policy is attached.
An object within the bucket is also a valid resource. The element value can be
either a single bucket or object ARN string or an array of ARNs.
"Action" lists which action(s) the policy controls. Its value can also be either
a string or array of S3 APIs. Each action is the API name prepended by "s3:".
"Principle" specifies which user(s) are granted or denied access to the bucket
resource. Its value can be a string or an object containing an array of users.
Valid users can be identified with an account ARN, account id, or user ARN.
There are also two optional bucket policy statement keys: Sid and Condition.
"Sid" stands for "statement id". If this key is not included, one will be
generated for the statement.
"Condition" lists the condition under which a statement will take affect.
The possibilities are as follows:
- ArnEquals
- ArnEqualsIfExists
- ArnLike
- ArnLikeIfExists
- ArnNotEquals
- ArnNotEqualsIfExists
- ArnNotLike
- ArnNotLikeIfExists
- BinaryEquals
- BinaryEqualsIfExists
- BinaryNotEquals
- BinaryNotEqualsIfExists
- Bool
- BoolIfExists
- DateEquals
- DateEqualsIfExists
- DateGreaterThan
- DateGreaterThanEquals
- DateGreaterThanEqualsIfExists
- DateGreaterThanIfExists
- DateLessThan
- DateLessThanEquals
- DateLessThanEqualsIfExists
- DateLessThanIfExists
- DateNotEquals
- DateNotEqualsIfExists
- IpAddress
- IpAddressIfExists
- NotIpAddress
- NotIpAddressIfExists
- Null
- NumericEquals
- NumericEqualsIfExists
- NumericGreaterThan
- NumericGreaterThanEquals
- NumericGreaterThanEqualsIfExists
- NumericGreaterThanIfExists
- NumericLessThan
- NumericLessThanEquals
- NumericLessThanEqualsIfExists
- NumericLessThanIfExists
- NumericNotEquals
- NumericNotEqualsIfExists
- StringEquals
- StringEqualsIfExists
- StringEqualsIgnoreCase
- StringEqualsIgnoreCaseIfExists
- StringLike
- StringLikeIfExists
- StringNotEquals
- StringNotEqualsIfExists
- StringNotEqualsIgnoreCase
- StringNotEqualsIgnoreCaseIfExists
- StringNotLike
- StringNotLikeIfExists
The value of the Condition key will be an object containing the desired
condition name as that key. The value of inner object can be a string, boolean,
number, or object, depending on the condition.
## Authorization with Multiple Access Control Mechanisms
In the case where multiple access control mechanisms (such as IAM policies,
bucket policies, and ACLs) refer to the same resource, the principle of
least-privilege is applied. Unless an action is explicitly allowed, access will
by default be denied. An explicit DENY in any policy will trump another
policy's ALLOW for an action. The request will only be allowed if at least one
policy specifies an ALLOW, and there is no overriding DENY.
The following diagram illustrates this logic:
![Access_Control_Authorization_Chart](./images/access_control_authorization.png)

View File

@ -178,7 +178,7 @@ Ruby
~~~~
`AWS SDK for Ruby - Version 2 <http://docs.aws.amazon.com/sdkforruby/api/>`__
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
.. code:: ruby
@ -239,7 +239,6 @@ Python
Client integration
.. code:: python
import boto3
client = boto3.client(
@ -254,7 +253,6 @@ Client integration
Full integration (with object mapping)
.. code:: python
import os
from botocore.utils import fix_s3_host
@ -295,51 +293,3 @@ Should force path-style requests even though v3 advertises it does by default.
$client->createBucket(array(
'Bucket' => 'bucketphp',
));
Go
~~
`AWS Go SDK <https://github.com/aws/aws-sdk-go>`__
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
.. code:: go
package main
import (
"context"
"fmt"
"log"
"os"
"time"
"github.com/aws/aws-sdk-go/aws"
"github.com/aws/aws-sdk-go/aws/endpoints"
"github.com/aws/aws-sdk-go/aws/session"
"github.com/aws/aws-sdk-go/service/s3"
)
func main() {
os.Setenv("AWS_ACCESS_KEY_ID", "accessKey1")
os.Setenv("AWS_SECRET_ACCESS_KEY", "verySecretKey1")
endpoint := "http://localhost:8000"
timeout := time.Duration(10) * time.Second
sess := session.Must(session.NewSession())
// Create a context with a timeout that will abort the upload if it takes
// more than the passed in timeout.
ctx, cancel := context.WithTimeout(context.Background(), timeout)
defer cancel()
svc := s3.New(sess, &aws.Config{
Region: aws.String(endpoints.UsEast1RegionID),
Endpoint: &endpoint,
})
out, err := svc.ListBucketsWithContext(ctx, &s3.ListBucketsInput{})
if err != nil {
log.Fatal(err)
} else {
fmt.Println(out)
}
}

View File

@ -14,7 +14,7 @@ Got an idea? Get started!
In order to contribute, please follow the `Contributing
Guidelines <https://github.com/scality/Guidelines/blob/master/CONTRIBUTING.md>`__.
If anything is unclear to you, reach out to us on
`forum <https://forum.zenko.io/>`__ or via a GitHub issue.
`slack <https://zenko-io.slack.com/>`__ or via a GitHub issue.
Don't write code? There are other ways to help!
-----------------------------------------------

View File

@ -1,7 +1,11 @@
Docker
======
.. _environment-variables:
- `Environment Variables <#environment-variables>`__
- `Tunables and setup tips <#tunables-and-setup-tips>`__
- `Examples for continuous integration with
Docker <#continuous-integration-with-docker-hosted CloudServer>`__
- `Examples for going in production with Docker <#in-production-with-docker-hosted CloudServer>`__
Environment Variables
---------------------
@ -11,27 +15,25 @@ S3DATA
S3DATA=multiple
^^^^^^^^^^^^^^^
This variable enables running CloudServer with multiple data backends, defined
Allows you to run Scality Zenko CloudServer with multiple data backends, defined
as regions.
When using multiple data backends, a custom ``locationConfig.json`` file is
mandatory. It will allow you to set custom regions. You will then need to
provide associated rest_endpoints for each custom region in your
``config.json`` file.
`Learn more about multiple backends configuration <../GETTING_STARTED/#location-configuration>`__
For multiple data backends, a custom locationConfig.json file is required.
This file enables you to set custom regions. You must provide associated
rest_endpoints for each custom region in config.json.
If you are using Scality RING endpoints, please refer to your customer
documentation.
`Learn more about multiple-backend configurations <GETTING_STARTED.html#location-configuration>`__
If you are using Scality RING endpoints, refer to your customer documentation.
Running CloudServer with an AWS S3-Hosted Backend
"""""""""""""""""""""""""""""""""""""""""""""""""
To run CloudServer with an S3 AWS backend, add a new section to the
``locationConfig.json`` file with the ``aws_s3`` location type:
Running it with an AWS S3 hosted backend
""""""""""""""""""""""""""""""""""""""""
To run CloudServer with an S3 AWS backend, you will have to add a new section
to your ``locationConfig.json`` file with the ``aws_s3`` location type:
.. code:: json
(...)
(...)
"awsbackend": {
"type": "aws_s3",
"details": {
@ -41,139 +43,126 @@ To run CloudServer with an S3 AWS backend, add a new section to the
"credentialsProfile": "aws_hosted_profile"
}
}
(...)
(...)
Edit your AWS credentials file to enable your preferred command-line tool.
This file must mention credentials for all backends in use. You can use
several profiles if multiple profiles are configured.
You will also have to edit your AWS credentials file to be able to use your
command line tool of choice. This file should mention credentials for all the
backends you're using. You can use several profiles when using multiple
profiles.
.. code:: json
[default]
aws_access_key_id=accessKey1
aws_secret_access_key=verySecretKey1
[aws_hosted_profile]
aws_access_key_id={{YOUR_ACCESS_KEY}}
aws_secret_access_key={{YOUR_SECRET_KEY}}
[default]
aws_access_key_id=accessKey1
aws_secret_access_key=verySecretKey1
[aws_hosted_profile]
aws_access_key_id={{YOUR_ACCESS_KEY}}
aws_secret_access_key={{YOUR_SECRET_KEY}}
As with locationConfig.json, the AWS credentials file must be mounted at
run time: ``-v ~/.aws/credentials:/root/.aws/credentials`` on Unix-like
systems (Linux, OS X, etc.), or
Just as you need to mount your locationConfig.json, you will need to mount your
AWS credentials file at run time:
``-v ~/.aws/credentials:/root/.aws/credentials`` on Linux, OS X, or Unix or
``-v C:\Users\USERNAME\.aws\credential:/root/.aws/credentials`` on Windows
.. note:: One account cannot copy to another account with a source and
destination on real AWS unless the account associated with the
accessKey/secretKey pairs used for the destination bucket has source
bucket access privileges. To enable this, update ACLs directly on AWS.
NOTE: One account can't copy to another account with a source and
destination on real AWS unless the account associated with the
access Key/secret Key pairs used for the destination bucket has rights
to get in the source bucket. ACL's would have to be updated
on AWS directly to enable this.
S3BACKEND
~~~~~~~~~
~~~~~~
S3BACKEND=file
^^^^^^^^^^^^^^
For stored file data to persist, you must mount Docker volumes
for both data and metadata. See :ref:`In Production with a Docker-Hosted CloudServer <in-production-w-a-Docker-hosted-cloudserver>`
^^^^^^^^^^^
When storing file data, for it to be persistent you must mount docker volumes
for both data and metadata. See `this section <#using-docker-volumes-in-production>`__
S3BACKEND=mem
^^^^^^^^^^^^^
This is ideal for testing: no data remains after the container is shut down.
^^^^^^^^^^
This is ideal for testing - no data will remain after container is shutdown.
ENDPOINT
~~~~~~~~
This variable specifies the endpoint. To direct CloudServer requests to
new.host.com, for example, specify the endpoint with:
This variable specifies your endpoint. If you have a domain such as
new.host.com, by specifying that here, you and your users can direct s3
server requests to new.host.com.
.. code-block:: shell
.. code:: shell
$ docker run -d --name cloudserver -p 8000:8000 -e ENDPOINT=new.host.com zenko/cloudserver
docker run -d --name s3server -p 8000:8000 -e ENDPOINT=new.host.com scality/s3server
.. note:: On Unix-like systems (Linux, OS X, etc.) edit /etc/hosts
to associate 127.0.0.1 with new.host.com.
REMOTE_MANAGEMENT_DISABLE
~~~~~~~~~~~~~~~~~~~~~~~~~
CloudServer is a part of `Zenko <https://www.zenko.io/>`__. When you run CloudServer standalone it will still try to connect to Orbit by default (browser-based graphical user interface for Zenko).
Setting this variable to true(1) will default to accessKey1 and verySecretKey1 for credentials and disable the automatic Orbit management:
.. code-block:: shell
$ docker run -d --name cloudserver -p 8000:8000 -e REMOTE_MANAGEMENT_DISABLE=1 zenko/cloudserver
Note: In your ``/etc/hosts`` file on Linux, OS X, or Unix with root
permissions, make sure to associate 127.0.0.1 with ``new.host.com``
SCALITY\_ACCESS\_KEY\_ID and SCALITY\_SECRET\_ACCESS\_KEY
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
These variables specify authentication credentials for an account named
“CustomAccount”.
"CustomAccount".
Set account credentials for multiple accounts by editing conf/authdata.json
(see below for further details). To specify one set for personal use, set these
environment variables:
You can set credentials for many accounts by editing
``conf/authdata.json`` (see below for further info), but if you just
want to specify one set of your own, you can use these environment
variables.
.. code-block:: shell
.. code:: shell
$ docker run -d --name cloudserver -p 8000:8000 -e SCALITY_ACCESS_KEY_ID=newAccessKey \
-e SCALITY_SECRET_ACCESS_KEY=newSecretKey zenko/cloudserver
docker run -d --name s3server -p 8000:8000 -e SCALITY_ACCESS_KEY_ID=newAccessKey
-e SCALITY_SECRET_ACCESS_KEY=newSecretKey scality/s3server
.. note:: This takes precedence over the contents of the authdata.json
file. The authdata.json file is ignored.
.. note:: The ACCESS_KEY and SECRET_KEY environment variables are
deprecated.
Note: Anything in the ``authdata.json`` file will be ignored. Note: The
old ``ACCESS_KEY`` and ``SECRET_KEY`` environment variables are now
deprecated
LOG\_LEVEL
~~~~~~~~~~
This variable changes the log level. There are three levels: info, debug,
and trace. The default is info. Debug provides more detailed logs, and trace
provides the most detailed logs.
This variable allows you to change the log level: info, debug or trace.
The default is info. Debug will give you more detailed logs and trace
will give you the most detailed.
.. code-block:: shell
.. code:: shell
$ docker run -d --name cloudserver -p 8000:8000 -e LOG_LEVEL=trace zenko/cloudserver
docker run -d --name s3server -p 8000:8000 -e LOG_LEVEL=trace scality/s3server
SSL
~~~
Set true, this variable runs CloudServer with SSL.
This variable set to true allows you to run S3 with SSL:
If SSL is set true:
**Note1**: You also need to specify the ENDPOINT environment variable.
**Note2**: In your ``/etc/hosts`` file on Linux, OS X, or Unix with root
permissions, make sure to associate 127.0.0.1 with ``<YOUR_ENDPOINT>``
* The ENDPOINT environment variable must also be specified.
* On Unix-like systems (Linux, OS X, etc.), 127.0.0.1 must be associated with
<YOUR_ENDPOINT> in /etc/hosts.
.. Warning:: Self-signed certs with a CA generated within the container are
suitable for testing purposes only. Clients cannot trust them, and they may
disappear altogether on a container upgrade. The best security practice for
production environments is to use an extra container, such as
haproxy/nginx/stunnel, for SSL/TLS termination and to pull certificates
from a mounted volume, limiting what an exploit on either component
can expose.
**Warning**: These certs, being self-signed (and the CA being generated
inside the container) will be untrusted by any clients, and could
disappear on a container upgrade. That's ok as long as it's for quick
testing. Also, best security practice for non-testing would be to use an
extra container to do SSL/TLS termination such as haproxy/nginx/stunnel
to limit what an exploit on either component could expose, as well as
certificates in a mounted volume
.. code:: shell
$ docker run -d --name cloudserver -p 8000:8000 -e SSL=TRUE -e ENDPOINT=<YOUR_ENDPOINT> \
zenko/cloudserver
docker run -d --name s3server -p 8000:8000 -e SSL=TRUE -e ENDPOINT=<YOUR_ENDPOINT>
scality/s3server
For more information about using ClousdServer with SSL, see `Using SSL <GETTING_STARTED.html#Using SSL>`__
More information about how to use S3 server with SSL
`here <https://s3.scality.com/v1.0/page/scality-with-ssl>`__
LISTEN\_ADDR
~~~~~~~~~~~~
This variable causes CloudServer and its data and metadata components to
listen on the specified address. This allows starting the data or metadata
servers as standalone services, for example.
This variable instructs the Zenko CloudServer, and its data and metadata
components to listen on the specified address. This allows starting the data
or metadata servers as standalone services, for example.
.. code:: shell
docker run -d --name s3server-data -p 9991:9991 -e LISTEN_ADDR=0.0.0.0
scality/s3server yarn run start_dataserver
scality/s3server npm run start_dataserver
DATA\_HOST and METADATA\_HOST
@ -185,8 +174,8 @@ Zenko CloudServer.
.. code:: shell
$ docker run -d --name cloudserver -e DATA_HOST=cloudserver-data \
-e METADATA_HOST=cloudserver-metadata zenko/cloudserver yarn run start_s3server
docker run -d --name s3server -e DATA_HOST=s3server-data
-e METADATA_HOST=s3server-metadata scality/s3server npm run start_s3server
REDIS\_HOST
~~~~~~~~~~~
@ -196,21 +185,19 @@ localhost.
.. code:: shell
$ docker run -d --name cloudserver -p 8000:8000 \
-e REDIS_HOST=my-redis-server.example.com zenko/cloudserver
docker run -d --name s3server -p 8000:8000
-e REDIS_HOST=my-redis-server.example.com scality/s3server
REDIS\_PORT
~~~~~~~~~~~
Use this variable to connect to the Redis cache server on a port other
than the default 6379.
Use this variable to connect to the redis cache server on another port than
the default 6379.
.. code:: shell
$ docker run -d --name cloudserver -p 8000:8000 \
-e REDIS_PORT=6379 zenko/cloudserver
.. _tunables-and-setup-tips:
docker run -d --name s3server -p 8000:8000
-e REDIS_PORT=6379 scality/s3server
Tunables and Setup Tips
-----------------------
@ -218,57 +205,60 @@ Tunables and Setup Tips
Using Docker Volumes
~~~~~~~~~~~~~~~~~~~~
CloudServer runs with a file backend by default, meaning that data is
stored inside the CloudServers Docker container.
Zenko CloudServer runs with a file backend by default.
For data and metadata to persist, data and metadata must be hosted in Docker
volumes outside the CloudServers Docker container. Otherwise, the data
and metadata are destroyed when the container is erased.
So, by default, the data is stored inside your Zenko CloudServer Docker
container.
.. code-block:: shell
However, if you want your data and metadata to persist, you **MUST** use
Docker volumes to host your data and metadata outside your Zenko CloudServer
Docker container. Otherwise, the data and metadata will be destroyed
when you erase the container.
$ docker run -­v $(pwd)/data:/usr/src/app/localData -­v $(pwd)/metadata:/usr/src/app/localMetadata \
-p 8000:8000 ­-d zenko/cloudserver
.. code:: shell
This command mounts the ./data host directory to the container
at /usr/src/app/localData and the ./metadata host directory to
the container at /usr/src/app/localMetaData.
docker run -­v $(pwd)/data:/usr/src/app/localData -­v $(pwd)/metadata:/usr/src/app/localMetadata
-p 8000:8000 ­-d scality/s3server
.. tip:: These host directories can be mounted to any accessible mount
point, such as /mnt/data and /mnt/metadata, for example.
This command mounts the host directory, ``./data``, into the container
at ``/usr/src/app/localData`` and the host directory, ``./metadata``, into
the container at ``/usr/src/app/localMetaData``. It can also be any host
mount point, like ``/mnt/data`` and ``/mnt/metadata``.
Adding, Modifying, or Deleting Accounts or Credentials
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Adding modifying or deleting accounts or users credentials
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1. Create a customized authdata.json file locally based on /conf/authdata.json.
2. Use `Docker volumes <https://docs.docker.com/storage/volumes/>`__
to override the default ``authdata.json`` through a Docker file mapping.
1. Create locally a customized ``authdata.json`` based on our ``/conf/authdata.json``.
2. Use `Docker
Volume <https://docs.docker.com/engine/tutorials/dockervolumes/>`__
to override the default ``authdata.json`` through a docker file mapping.
For example:
.. code-block:: shell
.. code:: shell
$ docker run -v $(pwd)/authdata.json:/usr/src/app/conf/authdata.json -p 8000:8000 -d \
zenko/cloudserver
docker run -v $(pwd)/authdata.json:/usr/src/app/conf/authdata.json -p 8000:8000 -d
scality/s3server
Specifying a Host Name
~~~~~~~~~~~~~~~~~~~~~~
Specifying your own host name
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To specify a host name (for example, s3.domain.name), provide your own
`config.json <https://github.com/scality/cloudserver/blob/master/config.json>`__
file using `Docker volumes <https://docs.docker.com/storage/volumes/>`__.
To specify a host name (e.g. s3.domain.name), you can provide your own
`config.json <https://github.com/scality/S3/blob/master/config.json>`__
using `Docker
Volume <https://docs.docker.com/engine/tutorials/dockervolumes/>`__.
First, add a new key-value pair to the restEndpoints section of your
config.json. Make the key the host name you want, and the value the default
location\_constraint for this endpoint.
First add a new key-value pair in the restEndpoints section of your
config.json. The key in the key-value pair should be the host name you
would like to add and the value is the default location\_constraint for
this endpoint.
For example, ``s3.example.com`` is mapped to ``us-east-1`` which is one
of the ``location_constraints`` listed in your locationConfig.json file
`here <https://github.com/scality/S3/blob/master/locationConfig.json>`__.
For more information about location configuration, see:
`GETTING STARTED <GETTING_STARTED.html#location-configuration>`__
More information about location configuration
`here <https://github.com/scality/S3/blob/master/README.md#location-configuration>`__
.. code:: json
@ -276,33 +266,33 @@ For more information about location configuration, see:
"localhost": "file",
"127.0.0.1": "file",
...
"cloudserver.example.com": "us-east-1"
"s3.example.com": "us-east-1"
},
Next, run CloudServer using a `Docker volume
<https://docs.docker.com/engine/tutorials/dockervolumes/>`__:
Then, run your Scality S3 Server using `Docker
Volume <https://docs.docker.com/engine/tutorials/dockervolumes/>`__:
.. code-block:: shell
.. code:: shell
$ docker run -v $(pwd)/config.json:/usr/src/app/config.json -p 8000:8000 -d zenko/cloudserver
docker run -v $(pwd)/config.json:/usr/src/app/config.json -p 8000:8000 -d scality/s3server
The local ``config.json`` file overrides the default one through a Docker
file mapping.
Your local ``config.json`` file will override the default one through a
docker file mapping.
Running as an Unprivileged User
Running as an unprivileged user
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
CloudServer runs as root by default.
Zenko CloudServer runs as root by default.
To change this, modify the dockerfile and specify a user before the
entry point.
You can change that by modifing the dockerfile and specifying a user
before the entrypoint.
The user must exist within the container, and must own the
/usr/src/app directory for CloudServer to run.
The user needs to exist within the container, and own the folder
**/usr/src/app** for Scality Zenko CloudServer to run properly.
For example, the following dockerfile lines can be modified:
For instance, you can modify these lines in the dockerfile:
.. code-block:: shell
.. code:: shell
...
&& groupadd -r -g 1001 scality \
@ -314,58 +304,54 @@ For example, the following dockerfile lines can be modified:
USER scality
ENTRYPOINT ["/usr/src/app/docker-entrypoint.sh"]
.. _continuous-integration-with-docker-hosted-cloudserver:
Continuous integration with Docker hosted CloudServer
-----------------------------------------------------
Continuous Integration with a Docker-Hosted CloudServer
-------------------------------------------------------
When you start the Docker Scality Zenko CloudServer image, you can adjust the
configuration of the Scality Zenko CloudServer instance by passing one or more
environment variables on the docker run command line.
When you start the Docker CloudServer image, you can adjust the
configuration of the CloudServer instance by passing one or more
environment variables on the ``docker run`` command line.
Sample ways to run it for CI are:
- With custom locations (one in-memory, one hosted on AWS), and custom
credentials mounted:
To run CloudServer for CI with custom locations (one in-memory,
one hosted on AWS), and custom credentials mounted:
.. code:: shell
.. code-block:: shell
docker run --name CloudServer -p 8000:8000
-v $(pwd)/locationConfig.json:/usr/src/app/locationConfig.json
-v $(pwd)/authdata.json:/usr/src/app/conf/authdata.json
-v ~/.aws/credentials:/root/.aws/credentials
-e S3DATA=multiple -e S3BACKEND=mem scality/s3server
$ docker run --name CloudServer -p 8000:8000 \
-v $(pwd)/locationConfig.json:/usr/src/app/locationConfig.json \
-v $(pwd)/authdata.json:/usr/src/app/conf/authdata.json \
-v ~/.aws/credentials:/root/.aws/credentials \
-e S3DATA=multiple -e S3BACKEND=mem zenko/cloudserver
- With custom locations, (one in-memory, one hosted on AWS, one file),
and custom credentials set as environment variables
(see `this section <#scality-access-key-id-and-scality-secret-access-key>`__):
To run CloudServer for CI with custom locations, (one in-memory, one
hosted on AWS, and one file), and custom credentials `set as environment
variables <GETTING_STARTED.html#scality-access-key-id-and-scality-secret-access-key>`__):
.. code:: shell
.. code-block:: shell
docker run --name CloudServer -p 8000:8000
-v $(pwd)/locationConfig.json:/usr/src/app/locationConfig.json
-v ~/.aws/credentials:/root/.aws/credentials
-v $(pwd)/data:/usr/src/app/localData -v $(pwd)/metadata:/usr/src/app/localMetadata
-e SCALITY_ACCESS_KEY_ID=accessKey1
-e SCALITY_SECRET_ACCESS_KEY=verySecretKey1
-e S3DATA=multiple -e S3BACKEND=mem scality/s3server
$ docker run --name CloudServer -p 8000:8000 \
-v $(pwd)/locationConfig.json:/usr/src/app/locationConfig.json \
-v ~/.aws/credentials:/root/.aws/credentials \
-v $(pwd)/data:/usr/src/app/localData -v $(pwd)/metadata:/usr/src/app/localMetadata \
-e SCALITY_ACCESS_KEY_ID=accessKey1 \
-e SCALITY_SECRET_ACCESS_KEY=verySecretKey1 \
-e S3DATA=multiple -e S3BACKEND=mem zenko/cloudserver
In production with Docker hosted CloudServer
--------------------------------------------
.. _in-production-w-a-Docker-hosted-cloudserver:
In production, we expect that data will be persistent, that you will use the
multiple backends capabilities of Zenko CloudServer, and that you will have a
custom endpoint for your local storage, and custom credentials for your local
storage:
In Production with a Docker-Hosted CloudServer
----------------------------------------------
.. code:: shell
Because data must persist in production settings, CloudServer offers
multiple-backend capabilities. This requires a custom endpoint
and custom credentials for local storage.
Customize these with:
.. code-block:: shell
$ docker run -d --name CloudServer \
-v $(pwd)/data:/usr/src/app/localData -v $(pwd)/metadata:/usr/src/app/localMetadata \
-v $(pwd)/locationConfig.json:/usr/src/app/locationConfig.json \
-v $(pwd)/authdata.json:/usr/src/app/conf/authdata.json \
-v ~/.aws/credentials:/root/.aws/credentials -e S3DATA=multiple \
-e ENDPOINT=custom.endpoint.com \
-p 8000:8000 ­-d zenko/cloudserver \
docker run -d --name CloudServer
-v $(pwd)/data:/usr/src/app/localData -v $(pwd)/metadata:/usr/src/app/localMetadata
-v $(pwd)/locationConfig.json:/usr/src/app/locationConfig.json
-v $(pwd)/authdata.json:/usr/src/app/conf/authdata.json
-v ~/.aws/credentials:/root/.aws/credentials -e S3DATA=multiple
-e ENDPOINT=custom.endpoint.com
-p 8000:8000 ­-d scality/s3server

View File

@ -1,221 +1,214 @@
Getting Started
===============
=================
.. figure:: ../res/scality-cloudserver-logo.png
:alt: Zenko CloudServer logo
Dependencies
------------
Building and running the Scality Zenko CloudServer requires node.js 10.x and
yarn v1.17.x. Up-to-date versions can be found at
`Nodesource <https://github.com/nodesource/distributions>`__.
|CircleCI| |Scality CI|
Installation
------------
1. Clone the source code
Dependencies
~~~~~~~~~~~~
.. code-block:: shell
Building and running the Scality Zenko CloudServer requires node.js 6.9.5 and
npm v3 . Up-to-date versions can be found at
`Nodesource <https://github.com/nodesource/distributions>`__.
$ git clone https://github.com/scality/cloudserver.git
2. Go to the cloudserver directory and use yarn to install the js dependencies.
.. code-block:: shell
$ cd cloudserver
$ yarn install
Running CloudServer with a File Backend
---------------------------------------
.. code-block:: shell
$ yarn start
This starts a Zenko CloudServer on port 8000. Two additional ports, 9990
and 9991, are also open locally for internal transfer of metadata and
data, respectively.
The default access key is accessKey1. The secret key is verySecretKey1.
By default, metadata files are saved in the localMetadata directory and
data files are saved in the localData directory in the local ./cloudserver
directory. These directories are pre-created within the repository. To
save data or metadata in different locations, you must specify them using
absolute paths. Thus, when starting the server:
.. code-block:: shell
$ mkdir -m 700 $(pwd)/myFavoriteDataPath
$ mkdir -m 700 $(pwd)/myFavoriteMetadataPath
$ export S3DATAPATH="$(pwd)/myFavoriteDataPath"
$ export S3METADATAPATH="$(pwd)/myFavoriteMetadataPath"
$ yarn start
Running CloudServer with Multiple Data Backends
-----------------------------------------------
.. code-block:: shell
$ export S3DATA='multiple'
$ yarn start
This starts a Zenko CloudServer on port 8000.
The default access key is accessKey1. The secret key is verySecretKey1.
With multiple backends, you can choose where each object is saved by setting
the following header with a location constraint in a PUT request:
.. code-block:: shell
'x-amz-meta-scal-location-constraint':'myLocationConstraint'
If no header is sent with a PUT object request, the buckets location
constraint determines where the data is saved. If the bucket has no
location constraint, the endpoint of the PUT request determines location.
See the Configuration_ section to set location constraints.
Run CloudServer with an In-Memory Backend
-----------------------------------------
.. code-block:: shell
$ yarn run mem_backend
This starts a Zenko CloudServer on port 8000.
The default access key is accessKey1. The secret key is verySecretKey1.
Run CloudServer with Vault User Management
------------------------------------------
Clone source code
~~~~~~~~~~~~~~~~~
.. code:: shell
export S3VAULT=vault
yarn start
git clone https://github.com/scality/S3.git
Note: Vault is proprietary and must be accessed separately.
This starts a Zenko CloudServer using Vault for user management.
Install js dependencies
~~~~~~~~~~~~~~~~~~~~~~~
Run CloudServer for Continuous Integration Testing or in Production with Docker
-------------------------------------------------------------------------------
Go to the ./S3 folder,
Run Cloudserver with `DOCKER <DOCKER.html>`__
.. code:: shell
Testing
~~~~~~~
npm install
Run unit tests with the command:
Run it with a file backend
--------------------------
.. code-block:: shell
.. code:: shell
$ yarn test
npm start
Run multiple-backend unit tests with:
This starts an Zenko CloudServer on port 8000. Two additional ports 9990 and
9991 are also open locally for internal transfer of metadata and data,
respectively.
.. code-block:: shell
The default access key is accessKey1 with a secret key of
verySecretKey1.
$ CI=true S3DATA=multiple yarn start
$ yarn run multiple_backend_test
By default the metadata files will be saved in the localMetadata
directory and the data files will be saved in the localData directory
within the ./S3 directory on your machine. These directories have been
pre-created within the repository. If you would like to save the data or
metadata in different locations of your choice, you must specify them
with absolute paths. So, when starting the server:
Run the linter with:
.. code:: shell
.. code-block:: shell
mkdir -m 700 $(pwd)/myFavoriteDataPath
mkdir -m 700 $(pwd)/myFavoriteMetadataPath
export S3DATAPATH="$(pwd)/myFavoriteDataPath"
export S3METADATAPATH="$(pwd)/myFavoriteMetadataPath"
npm start
$ yarn run lint
Run it with multiple data backends
----------------------------------
Running Functional Tests Locally
.. code:: shell
export S3DATA='multiple'
npm start
This starts an Zenko CloudServer on port 8000. The default access key is
accessKey1 with a secret key of verySecretKey1.
With multiple backends, you have the ability to choose where each object
will be saved by setting the following header with a locationConstraint
on a PUT request:
.. code:: shell
'x-amz-meta-scal-location-constraint':'myLocationConstraint'
If no header is sent with a PUT object request, the location constraint
of the bucket will determine where the data is saved. If the bucket has
no location constraint, the endpoint of the PUT request will be used to
determine location.
See the Configuration section below to learn how to set location
constraints.
Run it with an in-memory backend
--------------------------------
To pass AWS and Azure backend tests locally, modify
tests/locationConfig/locationConfigTests.json so that ``awsbackend``
specifies the bucketname of a bucket you have access to based on your
credentials, and modify ``azurebackend`` with details for your Azure account.
.. code:: shell
npm run mem_backend
This starts an Zenko CloudServer on port 8000. The default access key is
accessKey1 with a secret key of verySecretKey1.
Run it for continuous integration testing or in production with Docker
----------------------------------------------------------------------
`DOCKER <../DOCKER/>`__
Testing
-------
You can run the unit tests with the following command:
.. code:: shell
npm test
You can run the multiple backend unit tests with:
.. code:: shell
CI=true S3DATA=multiple npm start
npm run multiple_backend_test
You can run the linter with:
.. code:: shell
npm run lint
Running functional tests locally:
For the AWS backend and Azure backend tests to pass locally,
you must modify tests/locationConfigTests.json so that awsbackend
specifies a bucketname of a bucket you have access to based on
your credentials profile and modify "azurebackend" with details
for your Azure account.
The test suite requires additional tools, **s3cmd** and **Redis**
installed in the environment the tests are running in.
1. Install `s3cmd <http://s3tools.org/download>`__
- Install `s3cmd <http://s3tools.org/download>`__
- Install `redis <https://redis.io/download>`__ and start Redis.
- Add localCache section to your ``config.json``:
2. Install `redis <https://redis.io/download>`__ and start Redis.
::
3. Add localCache section to ``config.json``:
.. code:: json
"localCache": {
"localCache": {
"host": REDIS_HOST,
"port": REDIS_PORT
}
}
where ``REDIS_HOST`` is the Redis instance IP address (``"127.0.0.1"``
if Redis is running locally) and ``REDIS_PORT`` is the Redis instance
port (``6379`` by default)
where ``REDIS_HOST`` is your Redis instance IP address (``"127.0.0.1"``
if your Redis is running locally) and ``REDIS_PORT`` is your Redis
instance port (``6379`` by default)
4. Add the following to the local etc/hosts file:
- Add the following to the etc/hosts file on your machine:
.. code-block:: shell
.. code:: shell
127.0.0.1 bucketwebsitetester.s3-website-us-east-1.amazonaws.com
127.0.0.1 bucketwebsitetester.s3-website-us-east-1.amazonaws.com
5. Start Zenko CloudServer in memory and run the functional tests:
- Start the Zenko CloudServer in memory and run the functional tests:
.. code-block:: shell
.. code:: shell
$ CI=true yarn run mem_backend
$ CI=true yarn run ft_test
.. _Configuration:
CI=true npm run mem_backend
CI=true npm run ft_test
Configuration
-------------
There are three configuration files for Zenko CloudServer:
There are three configuration files for your Scality Zenko CloudServer:
* ``conf/authdata.json``, for authentication.
1. ``conf/authdata.json``, described above for authentication
* ``locationConfig.json``, to configure where data is saved.
2. ``locationConfig.json``, to set up configuration options for
* ``config.json``, for general configuration options.
where data will be saved
.. _location-configuration:
3. ``config.json``, for general configuration options
Location Configuration
~~~~~~~~~~~~~~~~~~~~~~
You must specify at least one locationConstraint in locationConfig.json
(or leave it as pre-configured).
You must specify at least one locationConstraint in your
locationConfig.json (or leave as pre-configured).
You must also specify 'us-east-1' as a locationConstraint. If you put a
bucket to an unknown endpoint and do not specify a locationConstraint in
the PUT bucket call, us-east-1 is used.
You must also specify 'us-east-1' as a locationConstraint so if you only
define one locationConstraint, that would be it. If you put a bucket to
an unknown endpoint and do not specify a locationConstraint in the put
bucket call, us-east-1 will be used.
For instance, the following locationConstraint saves data sent to
For instance, the following locationConstraint will save data sent to
``myLocationConstraint`` to the file backend:
.. code:: json
"myLocationConstraint": {
"type": "file",
"legacyAwsBehavior": false,
"details": {}
},
"myLocationConstraint": {
"type": "file",
"legacyAwsBehavior": false,
"details": {}
},
Each locationConstraint must include the ``type``, ``legacyAwsBehavior``,
and ``details`` keys. ``type`` indicates which backend is used for that
region. Supported backends are mem, file, and scality.``legacyAwsBehavior``
indicates whether the region behaves the same as the AWS S3 'us-east-1'
region. If the locationConstraint type is ``scality``, ``details`` must
contain connector information for sproxyd. If the locationConstraint type
is ``mem`` or ``file``, ``details`` must be empty.
Each locationConstraint must include the ``type``,
``legacyAwsBehavior``, and ``details`` keys. ``type`` indicates which
backend will be used for that region. Currently, mem, file, and scality
are the supported backends. ``legacyAwsBehavior`` indicates whether the
region will have the same behavior as the AWS S3 'us-east-1' region. If
the locationConstraint type is scality, ``details`` should contain
connector information for sproxyd. If the locationConstraint type is mem
or file, ``details`` should be empty.
Once locationConstraints is set in locationConfig.json, specify a default
locationConstraint for each endpoint.
Once you have your locationConstraints in your locationConfig.json, you
can specify a default locationConstraint for each of your endpoints.
For instance, the following sets the ``localhost`` endpoint to the
``myLocationConstraint`` data backend defined above:
@ -226,24 +219,26 @@ For instance, the following sets the ``localhost`` endpoint to the
"localhost": "myLocationConstraint"
},
To use an endpoint other than localhost for Zenko CloudServer, the endpoint
must be listed in ``restEndpoints``. Otherwise, if the server is running
with a:
If you would like to use an endpoint other than localhost for your
Scality Zenko CloudServer, that endpoint MUST be listed in your
``restEndpoints``. Otherwise if your server is running with a:
* **file backend**: The default location constraint is ``file``
* **memory backend**: The default location constraint is ``mem``
- **file backend**: your default location constraint will be ``file``
- **memory backend**: your default location constraint will be ``mem``
Endpoints
~~~~~~~~~
The Zenko CloudServer supports endpoints that are rendered in either:
Note that our Zenko CloudServer supports both:
* path style: http://myhostname.com/mybucket or
* hosted style: http://mybucket.myhostname.com
- path-style: http://myhostname.com/mybucket
- hosted-style: http://mybucket.myhostname.com
However, if an IP address is specified for the host, hosted-style requests
cannot reach the server. Use path-style requests in that case. For example,
if you are using the AWS SDK for JavaScript, instantiate your client like this:
However, hosted-style requests will not hit the server if you are using
an ip address for your host. So, make sure you are using path-style
requests in that case. For instance, if you are using the AWS SDK for
JavaScript, you would instantiate your client like this:
.. code:: js
@ -252,99 +247,87 @@ if you are using the AWS SDK for JavaScript, instantiate your client like this:
s3ForcePathStyle: true,
});
Setting Your Own Access and Secret Key Pairs
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Setting your own access key and secret key pairs
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Credentials can be set for many accounts by editing ``conf/authdata.json``,
but use the ``SCALITY_ACCESS_KEY_ID`` and ``SCALITY_SECRET_ACCESS_KEY``
environment variables to specify your own credentials.
_`scality-access-key-id-and-scality-secret-access-key`
You can set credentials for many accounts by editing
``conf/authdata.json`` but if you want to specify one set of your own
credentials, you can use ``SCALITY_ACCESS_KEY_ID`` and
``SCALITY_SECRET_ACCESS_KEY`` environment variables.
SCALITY\_ACCESS\_KEY\_ID and SCALITY\_SECRET\_ACCESS\_KEY
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
These variables specify authentication credentials for an account named
“CustomAccount”.
"CustomAccount".
.. note:: Anything in the ``authdata.json`` file is ignored.
Note: Anything in the ``authdata.json`` file will be ignored.
.. code-block:: shell
.. code:: shell
$ SCALITY_ACCESS_KEY_ID=newAccessKey SCALITY_SECRET_ACCESS_KEY=newSecretKey yarn start
SCALITY_ACCESS_KEY_ID=newAccessKey SCALITY_SECRET_ACCESS_KEY=newSecretKey npm start
.. _Using_SSL:
Using SSL
~~~~~~~~~
Scality with SSL
~~~~~~~~~~~~~~~~~~~~~~
To use https with your local CloudServer, you must set up
SSL certificates.
If you wish to use https with your local Zenko CloudServer, you need to set up
SSL certificates. Here is a simple guide of how to do it.
1. Deploy CloudServer using `our DockerHub page
<https://hub.docker.com/r/zenko/cloudserver/>`__ (run it with a file
backend).
Deploying Zenko CloudServer
^^^^^^^^^^^^^^^^^^^
.. Note:: If Docker is not installed locally, follow the
`instructions to install it for your distribution
<https://docs.docker.com/engine/installation/>`__
First, you need to deploy **Zenko CloudServer**. This can be done very easily
via `our **DockerHub**
page <https://hub.docker.com/r/scality/s3server/>`__ (you want to run it
with a file backend).
2. Update the CloudServer containers config
*Note:* *- If you don't have docker installed on your machine, here
are the `instructions to install it for your
distribution <https://docs.docker.com/engine/installation/>`__*
Add your certificates to your container. To do this,
#. exec inside the CloudServer container.
Updating your Zenko CloudServer container's config
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
#. Run ``$> docker ps`` to find the containers ID (the corresponding
image name is ``scality/cloudserver``.
#. Copy the corresponding container ID (``894aee038c5e`` in the present
example), and run:
You're going to add your certificates to your container. In order to do
so, you need to exec inside your Zenko CloudServer container. Run a
``$> docker ps`` and find your container's id (the corresponding image
name should be ``scality/s3server``. Copy the corresponding container id
(here we'll use ``894aee038c5e``, and run:
.. code-block:: shell
.. code:: sh
$> docker exec -it 894aee038c5e bash
$> docker exec -it 894aee038c5e bash
This puts you inside your container, using an interactive terminal.
You're now inside your container, using an interactive terminal :)
3. Generate the SSL key and certificates. The paths where the different
files are stored are defined after the ``-out`` option in each of the
following commands.
Generate SSL key and certificates
**********************************
#. Generate a private key for your certificate signing request (CSR):
There are 5 steps to this generation. The paths where the different
files are stored are defined after the ``-out`` option in each command
.. code-block:: shell
.. code:: sh
$> openssl genrsa -out ca.key 2048
# Generate a private key for your CSR
$> openssl genrsa -out ca.key 2048
# Generate a self signed certificate for your local Certificate Authority
$> openssl req -new -x509 -extensions v3_ca -key ca.key -out ca.crt -days 99999 -subj "/C=US/ST=Country/L=City/O=Organization/CN=scality.test"
#. Generate a self-signed certificate for your local certificate
authority (CA):
# Generate a key for Zenko CloudServer
$> openssl genrsa -out test.key 2048
# Generate a Certificate Signing Request for S3 Server
$> openssl req -new -key test.key -out test.csr -subj "/C=US/ST=Country/L=City/O=Organization/CN=*.scality.test"
# Generate a local-CA-signed certificate for S3 Server
$> openssl x509 -req -in test.csr -CA ca.crt -CAkey ca.key -CAcreateserial -out test.crt -days 99999 -sha256
.. code:: shell
Update Zenko CloudServer ``config.json``
**********************************
$> openssl req -new -x509 -extensions v3_ca -key ca.key -out ca.crt -days 99999 -subj "/C=US/ST=Country/L=City/O=Organization/CN=scality.test"
Add a ``certFilePaths`` section to ``./config.json`` with the
appropriate paths:
#. Generate a key for the CloudServer:
.. code:: shell
$> openssl genrsa -out test.key 2048
#. Generate a CSR for CloudServer:
.. code:: shell
$> openssl req -new -key test.key -out test.csr -subj "/C=US/ST=Country/L=City/O=Organization/CN=*.scality.test"
#. Generate a certificate for CloudServer signed by the local CA:
.. code:: shell
$> openssl x509 -req -in test.csr -CA ca.crt -CAkey ca.key -CAcreateserial -out test.crt -days 99999 -sha256
4. Update Zenko CloudServer ``config.json``. Add a ``certFilePaths``
section to ``./config.json`` with appropriate paths:
.. code:: json
.. code:: json
"certFilePaths": {
"key": "./test.key",
@ -352,36 +335,42 @@ SSL certificates.
"ca": "./ca.crt"
}
5. Run your container with the new config.
Run your container with the new config
****************************************
#. Exit the container by running ``$> exit``.
First, you need to exit your container. Simply run ``$> exit``. Then,
you need to restart your container. Normally, a simple
``$> docker restart s3server`` should do the trick.
#. Restart the container with ``$> docker restart cloudserver``.
Update your host config
^^^^^^^^^^^^^^^^^^^^^^^
6. Update the host configuration by adding s3.scality.test
to /etc/hosts:
Associates local IP addresses with hostname
*******************************************
.. code:: bash
In your ``/etc/hosts`` file on Linux, OS X, or Unix (with root
permissions), edit the line of localhost so it looks like this:
127.0.0.1 localhost s3.scality.test
::
7. Copy the local certificate authority (ca.crt in step 4) from your
container. Choose the path to save this file to (in the present
example, ``/root/ca.crt``), and run:
127.0.0.1 localhost s3.scality.test
.. code:: shell
Copy the local certificate authority from your container
*********************************************************
$> docker cp 894aee038c5e:/usr/src/app/ca.crt /root/ca.crt
In the above commands, it's the file named ``ca.crt``. Choose the path
you want to save this file at (here we chose ``/root/ca.crt``), and run
something like:
.. note:: Your container ID will be different, and your path to
ca.crt may be different.
.. code:: sh
Test the Config
^^^^^^^^^^^^^^^
$> docker cp 894aee038c5e:/usr/src/app/ca.crt /root/ca.crt
If aws-sdk is not installed, run ``$> yarn install aws-sdk``.
Test your config
^^^^^^^^^^^^^^^^^
Paste the following script into a file named "test.js":
If you do not have aws-sdk installed, run ``$> npm install aws-sdk``. In
a ``test.js`` file, paste the following script:
.. code:: js
@ -421,13 +410,8 @@ Paste the following script into a file named "test.js":
});
});
Now run this script with:
.. code::
$> nodejs test.js
On success, the script outputs ``SSL is cool!``.
Now run that script with ``$> nodejs test.js``. If all goes well, it
should output ``SSL is cool!``. Enjoy that added security!
.. |CircleCI| image:: https://circleci.com/gh/scality/S3.svg?style=svg

View File

@ -1,7 +1,6 @@
# S3 Healthcheck
Scality S3 exposes a healthcheck route `/live` on the port used
for the metrics (defaults to port 8002) which returns a
Scality S3 exposes a healthcheck route `/_/healthcheck` which returns a
response with HTTP code
- 200 OK

View File

@ -4,415 +4,479 @@ Integrations
High Availability
=================
`Docker Swarm <https://docs.docker.com/engine/swarm/>`__ is a clustering tool
developed by Docker for use with its containers. It can be used to start
services, which we define to ensure CloudServer's continuous availability to
end users. A swarm defines a manager and *n* workers among *n* + 1 servers.
This tutorial shows how to perform a basic setup with three servers, which
provides strong service resiliency, while remaining easy to use and
maintain. We will use NFS through Docker to share data and
`Docker swarm <https://docs.docker.com/engine/swarm/>`__ is a
clustering tool developped by Docker and ready to use with its
containers. It allows to start a service, which we define and use as a
means to ensure Zenko CloudServer's continuous availability to the end user.
Indeed, a swarm defines a manager and n workers among n+1 servers. We
will do a basic setup in this tutorial, with just 3 servers, which
already provides a strong service resiliency, whilst remaining easy to
do as an individual. We will use NFS through docker to share data and
metadata between the different servers.
Sections are labeled **On Server**, **On Clients**, or
**On All Machines**, referring respectively to NFS server, NFS clients, or
NFS server and clients. In the present example, the servers IP address is
**10.200.15.113** and the client IP addresses are **10.200.15.96** and
**10.200.15.97**
You will see that the steps of this tutorial are defined as **On
Server**, **On Clients**, **On All Machines**. This refers respectively
to NFS Server, NFS Clients, or NFS Server and Clients. In our example,
the IP of the Server will be **10.200.15.113**, while the IPs of the
Clients will be **10.200.15.96 and 10.200.15.97**
1. Install Docker (on All Machines)
Installing docker
-----------------
Docker 17.03.0-ce is used for this tutorial. Docker 1.12.6 and later will
likely work, but is not tested.
Any version from docker 1.12.6 onwards should work; we used Docker
17.03.0-ce for this tutorial.
* On Ubuntu 14.04
Install Docker CE for Ubuntu as `documented at Docker
<https://docs.docker.com/install/linux/docker-ce/ubuntu/>`__.
Install the aufs dependency as recommended by Docker. The required
commands are:
On All Machines
~~~~~~~~~~~~~~~
.. code:: sh
On Ubuntu 14.04
^^^^^^^^^^^^^^^
$> sudo apt-get update
$> sudo apt-get install linux-image-extra-$(uname -r) linux-image-extra-virtual
$> sudo apt-get install apt-transport-https ca-certificates curl software-properties-common
$> curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
$> sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"
$> sudo apt-get update
$> sudo apt-get install docker-ce
* On CentOS 7
Install Docker CE as `documented at Docker
<https://docs.docker.com/install/linux/docker-ce/centos/>`__.
The required commands are:
.. code:: sh
$> sudo yum install -y yum-utils
$> sudo yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo
$> sudo yum makecache fast
$> sudo yum install docker-ce
$> sudo systemctl start docker
2. Install NFS on Client(s)
NFS clients mount Docker volumes over the NFS servers shared folders.
If the NFS commons are installed, manual mounts are no longer needed.
* On Ubuntu 14.04
Install the NFS commons with apt-get:
.. code:: sh
$> sudo apt-get install nfs-common
* On CentOS 7
Install the NFS utils; then start required services:
.. code:: sh
$> yum install nfs-utils
$> sudo systemctl enable rpcbind
$> sudo systemctl enable nfs-server
$> sudo systemctl enable nfs-lock
$> sudo systemctl enable nfs-idmap
$> sudo systemctl start rpcbind
$> sudo systemctl start nfs-server
$> sudo systemctl start nfs-lock
$> sudo systemctl start nfs-idmap
3. Install NFS (on Server)
The NFS server hosts the data and metadata. The package(s) to install on it
differs from the package installed on the clients.
* On Ubuntu 14.04
Install the NFS server-specific package and the NFS commons:
.. code:: sh
$> sudo apt-get install nfs-kernel-server nfs-common
* On CentOS 7
Install the NFS utils and start the required services:
.. code:: sh
$> yum install nfs-utils
$> sudo systemctl enable rpcbind
$> sudo systemctl enable nfs-server
$> sudo systemctl enable nfs-lock
$> sudo systemctl enable nfs-idmap
$> sudo systemctl start rpcbind
$> sudo systemctl start nfs-server
$> sudo systemctl start nfs-lock
$> sudo systemctl start nfs-idmap
For both distributions:
#. Choose where shared data and metadata from the local
`CloudServer <http://www.zenko.io/cloudserver/>`__ shall be stored (The
present example uses /var/nfs/data and /var/nfs/metadata). Set permissions
for these folders for
sharing over NFS:
.. code:: sh
$> mkdir -p /var/nfs/data /var/nfs/metadata
$> chmod -R 777 /var/nfs/
#. The /etc/exports file configures network permissions and r-w-x permissions
for NFS access. Edit /etc/exports, adding the following lines:
.. code:: sh
/var/nfs/data 10.200.15.96(rw,sync,no_root_squash) 10.200.15.97(rw,sync,no_root_squash)
/var/nfs/metadata 10.200.15.96(rw,sync,no_root_squash) 10.200.15.97(rw,sync,no_root_squash)
Ubuntu applies the no\_subtree\_check option by default, so both
folders are declared with the same permissions, even though theyre in
the same tree.
#. Export this new NFS table:
.. code:: sh
$> sudo exportfs -a
#. Edit the ``MountFlags`` option in the Docker config in
/lib/systemd/system/docker.service to enable NFS mount from Docker volumes
on other machines:
.. code:: sh
MountFlags=shared
#. Restart the NFS server and Docker daemons to apply these changes.
* On Ubuntu 14.04
.. code:: sh
$> sudo service nfs-kernel-server restart
$> sudo service docker restart
* On CentOS 7
.. code:: sh
$> sudo systemctl restart nfs-server
$> sudo systemctl daemon-reload
$> sudo systemctl restart docker
4. Set Up a Docker Swarm
* On all machines and distributions:
Set up the Docker volumes to be mounted to the NFS server for CloudServers
data and metadata storage. The following commands must be replicated on all
machines:
.. code:: sh
$> docker volume create --driver local --opt type=nfs --opt o=addr=10.200.15.113,rw --opt device=:/var/nfs/data --name data
$> docker volume create --driver local --opt type=nfs --opt o=addr=10.200.15.113,rw --opt device=:/var/nfs/metadata --name metadata
There is no need to ``docker exec`` these volumes to mount them: the
Docker Swarm manager does this when the Docker service is started.
* On a server:
To start a Docker service on a Docker Swarm cluster, initialize the cluster
(that is, define a manager), prompt workers/nodes to join in, and then start
the service.
Initialize the swarm cluster, and review its response:
.. code:: sh
$> docker swarm init --advertise-addr 10.200.15.113
Swarm initialized: current node (db2aqfu3bzfzzs9b1kfeaglmq) is now a manager.
To add a worker to this swarm, run the following command:
docker swarm join \
--token SWMTKN-1-5yxxencrdoelr7mpltljn325uz4v6fe1gojl14lzceij3nujzu-2vfs9u6ipgcq35r90xws3stka \
10.200.15.113:2377
To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.
* On clients:
Copy and paste the command provided by your Docker Swarm init. A successful
request/response will resemble:
.. code:: sh
$> docker swarm join --token SWMTKN-1-5yxxencrdoelr7mpltljn325uz4v6fe1gojl14lzceij3nujzu-2vfs9u6ipgcq35r90xws3stka 10.200.15.113:2377
This node joined a swarm as a worker.
Set Up Docker Swarm on Clients on a Server
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Start the service on the Swarm cluster.
The docker website has `solid
documentation <https://docs.docker.com/engine/installation/linux/ubuntu/>`__.
We have chosen to install the aufs dependency, as recommended by Docker.
Here are the required commands:
.. code:: sh
$> docker service create --name s3 --replicas 1 --mount type=volume,source=data,target=/usr/src/app/localData --mount type=volume,source=metadata,target=/usr/src/app/localMetadata -p 8000:8000 scality/cloudserver
$> sudo apt-get update
$> sudo apt-get install linux-image-extra-$(uname -r) linux-image-extra-virtual
$> sudo apt-get install apt-transport-https ca-certificates curl software-properties-common
$> curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo apt-key add -
$> sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"
$> sudo apt-get update
$> sudo apt-get install docker-ce
On a successful installation, ``docker service ls`` returns the following
output:
On CentOS 7
^^^^^^^^^^^
The docker website has `solid
documentation <https://docs.docker.com/engine/installation/linux/centos/>`__.
Here are the required commands:
.. code:: sh
$> sudo yum install -y yum-utils
$> sudo yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce.repo
$> sudo yum makecache fast
$> sudo yum install docker-ce
$> sudo systemctl start docker
Configure NFS
-------------
On Clients
~~~~~~~~~~
Your NFS Clients will mount Docker volumes over your NFS Server's shared
folders. Hence, you don't have to mount anything manually, you just have
to install the NFS commons:
On Ubuntu 14.04
^^^^^^^^^^^^^^^
Simply install the NFS commons:
.. code:: sh
$> sudo apt-get install nfs-common
On CentOS 7
^^^^^^^^^^^
Install the NFS utils, and then start the required services:
.. code:: sh
$> yum install nfs-utils
$> sudo systemctl enable rpcbind
$> sudo systemctl enable nfs-server
$> sudo systemctl enable nfs-lock
$> sudo systemctl enable nfs-idmap
$> sudo systemctl start rpcbind
$> sudo systemctl start nfs-server
$> sudo systemctl start nfs-lock
$> sudo systemctl start nfs-idmap
On Server
~~~~~~~~~
Your NFS Server will be the machine to physically host the data and
metadata. The package(s) we will install on it is slightly different
from the one we installed on the clients.
On Ubuntu 14.04
^^^^^^^^^^^^^^^
Install the NFS server specific package and the NFS commons:
.. code:: sh
$> sudo apt-get install nfs-kernel-server nfs-common
On CentOS 7
^^^^^^^^^^^
Same steps as with the client: install the NFS utils and start the
required services:
.. code:: sh
$> yum install nfs-utils
$> sudo systemctl enable rpcbind
$> sudo systemctl enable nfs-server
$> sudo systemctl enable nfs-lock
$> sudo systemctl enable nfs-idmap
$> sudo systemctl start rpcbind
$> sudo systemctl start nfs-server
$> sudo systemctl start nfs-lock
$> sudo systemctl start nfs-idmap
On Ubuntu 14.04 and CentOS 7
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Choose where your shared data and metadata from your local `Zenko CloudServer
<http://www.zenko.io/cloudserver/>`__ will be stored.
We chose to go with /var/nfs/data and /var/nfs/metadata. You also need
to set proper sharing permissions for these folders as they'll be shared
over NFS:
.. code:: sh
$> mkdir -p /var/nfs/data /var/nfs/metadata
$> chmod -R 777 /var/nfs/
Now you need to update your **/etc/exports** file. This is the file that
configures network permissions and rwx permissions for NFS access. By
default, Ubuntu applies the no\_subtree\_check option, so we declared
both folders with the same permissions, even though they're in the same
tree:
.. code:: sh
$> sudo vim /etc/exports
In this file, add the following lines:
.. code:: sh
/var/nfs/data 10.200.15.96(rw,sync,no_root_squash) 10.200.15.97(rw,sync,no_root_squash)
/var/nfs/metadata 10.200.15.96(rw,sync,no_root_squash) 10.200.15.97(rw,sync,no_root_squash)
Export this new NFS table:
.. code:: sh
$> sudo exportfs -a
Eventually, you need to allow for NFS mount from Docker volumes on other
machines. You need to change the Docker config in
**/lib/systemd/system/docker.service**:
.. code:: sh
$> sudo vim /lib/systemd/system/docker.service
In this file, change the **MountFlags** option:
.. code:: sh
MountFlags=shared
Now you just need to restart the NFS server and docker daemons so your
changes apply.
On Ubuntu 14.04
^^^^^^^^^^^^^^^
Restart your NFS Server and docker services:
.. code:: sh
$> sudo service nfs-kernel-server restart
$> sudo service docker restart
On CentOS 7
^^^^^^^^^^^
Restart your NFS Server and docker daemons:
.. code:: sh
$> sudo systemctl restart nfs-server
$> sudo systemctl daemon-reload
$> sudo systemctl restart docker
Set up your Docker Swarm service
--------------------------------
On All Machines
~~~~~~~~~~~~~~~
On Ubuntu 14.04 and CentOS 7
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
We will now set up the Docker volumes that will be mounted to the NFS
Server and serve as data and metadata storage for Zenko CloudServer. These two
commands have to be replicated on all machines:
.. code:: sh
$> docker volume create --driver local --opt type=nfs --opt o=addr=10.200.15.113,rw --opt device=:/var/nfs/data --name data
$> docker volume create --driver local --opt type=nfs --opt o=addr=10.200.15.113,rw --opt device=:/var/nfs/metadata --name metadata
There is no need to ""docker exec" these volumes to mount them: the
Docker Swarm manager will do it when the Docker service will be started.
On Server
^^^^^^^^^
To start a Docker service on a Docker Swarm cluster, you first have to
initialize that cluster (i.e.: define a manager), then have the
workers/nodes join in, and then start the service. Initialize the swarm
cluster, and look at the response:
.. code:: sh
$> docker swarm init --advertise-addr 10.200.15.113
Swarm initialized: current node (db2aqfu3bzfzzs9b1kfeaglmq) is now a manager.
To add a worker to this swarm, run the following command:
docker swarm join \
--token SWMTKN-1-5yxxencrdoelr7mpltljn325uz4v6fe1gojl14lzceij3nujzu-2vfs9u6ipgcq35r90xws3stka \
10.200.15.113:2377
To add a manager to this swarm, run 'docker swarm join-token manager' and follow the instructions.
On Clients
^^^^^^^^^^
Simply copy/paste the command provided by your docker swarm init. When
all goes well, you'll get something like this:
.. code:: sh
$> docker swarm join --token SWMTKN-1-5yxxencrdoelr7mpltljn325uz4v6fe1gojl14lzceij3nujzu-2vfs9u6ipgcq35r90xws3stka 10.200.15.113:2377
This node joined a swarm as a worker.
On Server
^^^^^^^^^
Start the service on your swarm cluster!
.. code:: sh
$> docker service create --name s3 --replicas 1 --mount type=volume,source=data,target=/usr/src/app/localData --mount type=volume,source=metadata,target=/usr/src/app/localMetadata -p 8000:8000 scality/s3server
If you run a docker service ls, you should have the following output:
.. code:: sh
$> docker service ls
ID NAME MODE REPLICAS IMAGE
ocmggza412ft s3 replicated 1/1 scality/cloudserver:latest
ocmggza412ft s3 replicated 1/1 scality/s3server:latest
If the service does not start, consider disabling apparmor/SELinux.
If your service won't start, consider disabling apparmor/SELinux.
Testing the High-Availability CloudServer
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Testing your High Availability S3Server
---------------------------------------
On all machines (client/server) and distributions (Ubuntu and CentOS),
determine where CloudServer is running using ``docker ps``. CloudServer can
operate on any node of the Swarm cluster, manager or worker. When you find
it, you can kill it with ``docker stop <container id>``. It will respawn
on a different node. Now, if one server falls, or if Docker stops
unexpectedly, the end user will still be able to access your the local CloudServer.
Troubleshooting
On All Machines
~~~~~~~~~~~~~~~
To troubleshoot the service, run:
On Ubuntu 14.04 and CentOS 7
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Try to find out where your Scality Zenko CloudServer is actually running using
the **docker ps** command. It can be on any node of the swarm cluster,
manager or worker. When you find it, you can kill it, with **docker stop
<container id>** and you'll see it respawn on a different node of the
swarm cluster. Now you see, if one of your servers falls, or if docker
stops unexpectedly, your end user will still be able to access your
local Zenko CloudServer.
Troubleshooting
---------------
To troubleshoot the service you can run:
.. code:: sh
$> docker service ps s3docker service ps s3
ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR
0ar81cw4lvv8chafm8pw48wbc s3.1 scality/cloudserver localhost.localdomain.localdomain Running Running 7 days ago
cvmf3j3bz8w6r4h0lf3pxo6eu \_ s3.1 scality/cloudserver localhost.localdomain.localdomain Shutdown Failed 7 days ago "task: non-zero exit (137)"
0ar81cw4lvv8chafm8pw48wbc s3.1 scality/s3server localhost.localdomain.localdomain Running Running 7 days ago
cvmf3j3bz8w6r4h0lf3pxo6eu \_ s3.1 scality/s3server localhost.localdomain.localdomain Shutdown Failed 7 days ago "task: non-zero exit (137)"
If the error is truncated, view the error in detail by inspecting the
Docker task ID:
If the error is truncated it is possible to have a more detailed view of
the error by inspecting the docker task ID:
.. code:: sh
$> docker inspect cvmf3j3bz8w6r4h0lf3pxo6eu
Off you go!
~~~~~~~~~~~
-----------
Let us know what you use this functionality for, and if you'd like any
specific developments around it. Or, even better: come and contribute to
our `Github repository <https://github.com/scality/s3/>`__! We look
forward to meeting you!
Let us know how you use this and if you'd like any specific developments
around it. Even better: come and contribute to our `Github repository
<https://github.com/scality/s3/>`__! We look forward to meeting you!
S3FS
====
You can export buckets as a filesystem with s3fs on CloudServer.
Export your buckets as a filesystem with s3fs on top of Zenko CloudServer
`s3fs <https://github.com/s3fs-fuse/s3fs-fuse>`__ is an open source
tool, available both on Debian and RedHat distributions, that enables
you to mount an S3 bucket on a filesystem-like backend. This tutorial uses
an Ubuntu 14.04 host to deploy and use s3fs over CloudServer.
tool that allows you to mount an S3 bucket on a filesystem-like backend.
It is available both on Debian and RedHat distributions. For this
tutorial, we used an Ubuntu 14.04 host to deploy and use s3fs over
Scality's Zenko CloudServer.
Deploying Zenko CloudServer with SSL
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
----------------------------
First, deploy CloudServer with a file backend using `our DockerHub page
<https://hub.docker.com/r/zenko/cloudserver>`__.
First, you need to deploy **Zenko CloudServer**. This can be done very easily
via `our DockerHub
page <https://hub.docker.com/r/scality/s3server/>`__ (you want to run it
with a file backend).
.. note::
*Note:* *- If you don't have docker installed on your machine, here
are the `instructions to install it for your
distribution <https://docs.docker.com/engine/installation/>`__*
If Docker is not installed on your machine, follow
`these instructions <https://docs.docker.com/engine/installation/>`__
to install it for your distribution.
You also necessarily have to set up SSL with Zenko CloudServer to use s3fs. We
have a nice
`tutorial <https://s3.scality.com/v1.0/page/scality-with-ssl>`__ to help
you do it.
You must also set up SSL with CloudServer to use s3fs. See `Using SSL
<./GETTING_STARTED#Using_SSL>`__ for instructions.
s3fs Setup
~~~~~~~~~~
s3fs setup
----------
Installing s3fs
---------------
~~~~~~~~~~~~~~~
Follow the instructions in the s3fs `README
<https://github.com/s3fs-fuse/s3fs-fuse/blob/master/README.md#installation-from-pre-built-packages>`__,
Check that s3fs is properly installed. A version check should return
a response resembling:
s3fs has quite a few dependencies. As explained in their
`README <https://github.com/s3fs-fuse/s3fs-fuse/blob/master/README.md#installation>`__,
the following commands should install everything for Ubuntu 14.04:
.. code:: sh
$> s3fs --version
$> sudo apt-get install automake autotools-dev g++ git libcurl4-gnutls-dev
$> sudo apt-get install libfuse-dev libssl-dev libxml2-dev make pkg-config
Now you want to install s3fs per se:
.. code:: sh
$> git clone https://github.com/s3fs-fuse/s3fs-fuse.git
$> cd s3fs-fuse
$> ./autogen.sh
$> ./configure
$> make
$> sudo make install
Check that s3fs is properly installed by checking its version. it should
answer as below:
.. code:: sh
$> s3fs --version
Amazon Simple Storage Service File System V1.80(commit:d40da2c) with OpenSSL
Copyright (C) 2010 Randy Rizun <rrizun@gmail.com>
License GPL2: GNU GPL version 2 <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.
Configuring s3fs
----------------
~~~~~~~~~~~~~~~~
s3fs expects you to provide it with a password file. Our file is
``/etc/passwd-s3fs``. The structure for this file is
``ACCESSKEYID:SECRETKEYID``, so, for CloudServer, you can run:
``ACCESSKEYID:SECRETKEYID``, so, for S3Server, you can run:
.. code:: sh
$> echo 'accessKey1:verySecretKey1' > /etc/passwd-s3fs
$> chmod 600 /etc/passwd-s3fs
Using CloudServer with s3fs
---------------------------
Using Zenko CloudServer with s3fs
------------------------
1. Use /mnt/tests3fs as a mount point.
First, you're going to need a mountpoint; we chose ``/mnt/tests3fs``:
.. code:: sh
.. code:: sh
$> mkdir /mnt/tests3fs
2. Create a bucket on your local CloudServer. In the present example it is
named “tests3fs”.
Then, you want to create a bucket on your local Zenko CloudServer; we named it
``tests3fs``:
.. code:: sh
.. code:: sh
$> s3cmd mb s3://tests3fs
3. Mount the bucket to your mount point with s3fs:
*Note:* *- If you've never used s3cmd with our Zenko CloudServer, our README
provides you with a `recommended
config <https://github.com/scality/S3/blob/master/README.md#s3cmd>`__*
.. code:: sh
Now you can mount your bucket to your mountpoint with s3fs:
.. code:: sh
$> s3fs tests3fs /mnt/tests3fs -o passwd_file=/etc/passwd-s3fs -o url="https://s3.scality.test:8000/" -o use_path_request_style
The structure of this command is:
``s3fs BUCKET_NAME PATH/TO/MOUNTPOINT -o OPTIONS``. Of these mandatory
options:
*If you're curious, the structure of this command is*
``s3fs BUCKET_NAME PATH/TO/MOUNTPOINT -o OPTIONS``\ *, and the
options are mandatory and serve the following purposes:
* ``passwd_file``\ *: specifiy path to password file;
* ``url``\ *: specify the hostname used by your SSL provider;
* ``use_path_request_style``\ *: force path style (by default, s3fs
uses subdomains (DNS style)).*
* ``passwd_file`` specifies the path to the password file.
* ``url`` specifies the host name used by your SSL provider.
* ``use_path_request_style`` forces the path style (by default,
s3fs uses DNS-style subdomains).
| From now on, you can either add files to your mountpoint, or add
objects to your bucket, and they'll show in the other.
| For example, let's' create two files, and then a directory with a file
in our mountpoint:
Once the bucket is mounted, files added to the mount point or
objects added to the bucket will appear in both locations.
.. code:: sh
Example
-------
$> touch /mnt/tests3fs/file1 /mnt/tests3fs/file2
$> mkdir /mnt/tests3fs/dir1
$> touch /mnt/tests3fs/dir1/file3
Create two files, and then a directory with a file in our mount point:
Now, I can use s3cmd to show me what is actually in S3Server:
.. code:: sh
.. code:: sh
$> touch /mnt/tests3fs/file1 /mnt/tests3fs/file2
$> mkdir /mnt/tests3fs/dir1
$> touch /mnt/tests3fs/dir1/file3
$> s3cmd ls -r s3://tests3fs
Now, use s3cmd to show what is in CloudServer:
2017-02-28 17:28 0 s3://tests3fs/dir1/
2017-02-28 17:29 0 s3://tests3fs/dir1/file3
2017-02-28 17:28 0 s3://tests3fs/file1
2017-02-28 17:28 0 s3://tests3fs/file2
.. code:: sh
$> s3cmd ls -r s3://tests3fs
2017-02-28 17:28 0 s3://tests3fs/dir1/
2017-02-28 17:29 0 s3://tests3fs/dir1/file3
2017-02-28 17:28 0 s3://tests3fs/file1
2017-02-28 17:28 0 s3://tests3fs/file2
Now you can enjoy a filesystem view on your local CloudServer.
Now you can enjoy a filesystem view on your local Zenko CloudServer!
Duplicity
=========
How to back up your files with CloudServer.
How to backup your files with Zenko CloudServer.
Installing Duplicity and its Dependencies
Installing
-----------
Installing Duplicity and its dependencies
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
To install `Duplicity <http://duplicity.nongnu.org/>`__,
go to `this site <https://code.launchpad.net/duplicity/0.7-series>`__.
Download the latest tarball. Decompress it and follow the instructions
in the README.
.. code:: sh
$> tar zxvf duplicity-0.7.11.tar.gz
$> cd duplicity-0.7.11
$> python setup.py install
You may receive error messages indicating the need to install some or all
of the following dependencies:
Second, you want to install
`Duplicity <http://duplicity.nongnu.org/index.html>`__. You have to
download `this
tarball <https://code.launchpad.net/duplicity/0.7-series/0.7.11/+download/duplicity-0.7.11.tar.gz>`__,
decompress it, and then checkout the README inside, which will give you
a list of dependencies to install. If you're using Ubuntu 14.04, this is
your lucky day: here is a lazy step by step install.
.. code:: sh
@ -420,20 +484,30 @@ of the following dependencies:
$> apt-get install python-dev python-pip python-lockfile
$> pip install -U boto
Testing the Installation
------------------------
Then you want to actually install Duplicity:
1. Check that CloudServer is running. Run ``$> docker ps``. You should
see one container named ``scality/cloudserver``. If you do not, run
``$> docker start cloudserver`` and check again.
.. code:: sh
$> tar zxvf duplicity-0.7.11.tar.gz
$> cd duplicity-0.7.11
$> python setup.py install
2. Duplicity uses a module called “Boto” to send requests to S3. Boto
requires a configuration file located in ``/etc/boto.cfg`` to store
your credentials and preferences. A minimal configuration
you can fine tune `following these instructions
<http://boto.cloudhackers.com/en/latest/getting_started.html>`__ is
shown here:
Using
------
Testing your installation
~~~~~~~~~~~~~~~~~~~~~~~~~~~
First, we're just going to quickly check that Zenko CloudServer is actually
running. To do so, simply run ``$> docker ps`` . You should see one
container named ``scality/s3server``. If that is not the case, try
``$> docker start s3server``, and check again.
Secondly, as you probably know, Duplicity uses a module called **Boto**
to send requests to S3. Boto requires a configuration file located in
**``/etc/boto.cfg``** to have your credentials and preferences. Here is
a minimalistic config `that you can finetune following these
instructions <http://boto.cloudhackers.com/en/latest/getting_started.html>`__.
::
@ -447,51 +521,54 @@ Testing the Installation
# If using SSL, unmute and provide absolute path to local CA certificate
# ca_certificates_file = /absolute/path/to/ca.crt
.. note:: To set up SSL with CloudServer, check out our `Using SSL
<./GETTING_STARTED#Using_SSL>`__ in GETTING STARTED.
*Note:* *If you want to set up SSL with Zenko CloudServer, check out our
`tutorial <http://link/to/SSL/tutorial>`__*
3. At this point all requirements to run CloudServer as a backend to Duplicity
have been met. A local folder/file should back up to the local S3.
Try it with the decompressed Duplicity folder:
At this point, we've met all the requirements to start running Zenko CloudServer
as a backend to Duplicity. So we should be able to back up a local
folder/file to local S3. Let's try with the duplicity decompressed
folder:
.. code:: sh
$> duplicity duplicity-0.7.11 "s3://127.0.0.1:8000/testbucket/"
.. note:: Duplicity will prompt for a symmetric encryption passphrase.
Save it carefully, as you will need it to recover your data.
Alternatively, you can add the ``--no-encryption`` flag
and the data will be stored plain.
*Note:* *Duplicity will prompt you for a symmetric encryption
passphrase. Save it somewhere as you will need it to recover your
data. Alternatively, you can also add the ``--no-encryption`` flag
and the data will be stored plain.*
If this command is successful, you will receive an output resembling:
If this command is succesful, you will get an output looking like this:
.. code:: sh
::
--------------[ Backup Statistics ]--------------
StartTime 1486486547.13 (Tue Feb 7 16:55:47 2017)
EndTime 1486486547.40 (Tue Feb 7 16:55:47 2017)
ElapsedTime 0.27 (0.27 seconds)
SourceFiles 388
SourceFileSize 6634529 (6.33 MB)
NewFiles 388
NewFileSize 6634529 (6.33 MB)
DeletedFiles 0
ChangedFiles 0
ChangedFileSize 0 (0 bytes)
ChangedDeltaSize 0 (0 bytes)
DeltaEntries 388
RawDeltaSize 6392865 (6.10 MB)
TotalDestinationSizeChange 2003677 (1.91 MB)
Errors 0
-------------------------------------------------
--------------[ Backup Statistics ]--------------
StartTime 1486486547.13 (Tue Feb 7 16:55:47 2017)
EndTime 1486486547.40 (Tue Feb 7 16:55:47 2017)
ElapsedTime 0.27 (0.27 seconds)
SourceFiles 388
SourceFileSize 6634529 (6.33 MB)
NewFiles 388
NewFileSize 6634529 (6.33 MB)
DeletedFiles 0
ChangedFiles 0
ChangedFileSize 0 (0 bytes)
ChangedDeltaSize 0 (0 bytes)
DeltaEntries 388
RawDeltaSize 6392865 (6.10 MB)
TotalDestinationSizeChange 2003677 (1.91 MB)
Errors 0
-------------------------------------------------
Congratulations! You can now back up to your local S3 through Duplicity.
Congratulations! You can now backup to your local S3 through duplicity
:)
Automating Backups
------------------
Automating backups
~~~~~~~~~~~~~~~~~~~
The easiest way to back up files periodically is to write a bash script
and add it to your crontab. A suggested script follows.
Now you probably want to back up your files periodically. The easiest
way to do this is to write a bash script and add it to your crontab.
Here is my suggestion for such a file:
.. code:: sh
@ -500,33 +577,33 @@ and add it to your crontab. A suggested script follows.
# Export your passphrase so you don't have to type anything
export PASSPHRASE="mypassphrase"
# To use a GPG key, put it here and uncomment the line below
# If you want to use a GPG Key, put it here and unmute the line below
#GPG_KEY=
# Define your backup bucket, with localhost specified
DEST="s3://127.0.0.1:8000/testbucketcloudserver/"
DEST="s3://127.0.0.1:8000/testbuckets3server/"
# Define the absolute path to the folder to back up
# Define the absolute path to the folder you want to backup
SOURCE=/root/testfolder
# Set to "full" for full backups, and "incremental" for incremental backups
# Warning: you must perform one full backup befor you can perform
# Warning: you have to perform one full backup befor you can perform
# incremental ones on top of it
FULL=incremental
# How long to keep backups. If you don't want to delete old backups, keep
# this value empty; otherwise, the syntax is "1Y" for one year, "1M" for
# one month, "1D" for one day.
# How long to keep backups for; if you don't want to delete old
# backups, keep empty; otherwise, syntax is "1Y" for one year, "1M"
# for one month, "1D" for one day
OLDER_THAN="1Y"
# is_running checks whether Duplicity is currently completing a task
# is_running checks whether duplicity is currently completing a task
is_running=$(ps -ef | grep duplicity | grep python | wc -l)
# If Duplicity is already completing a task, this will not run
# If duplicity is already completing a task, this will simply not run
if [ $is_running -eq 0 ]; then
echo "Backup for ${SOURCE} started"
# To delete backups older than a certain time, do it here
# If you want to delete backups older than a certain time, we do it here
if [ "$OLDER_THAN" != "" ]; then
echo "Removing backups older than ${OLDER_THAN}"
duplicity remove-older-than ${OLDER_THAN} ${DEST}
@ -549,17 +626,17 @@ and add it to your crontab. A suggested script follows.
# Forget the passphrase...
unset PASSPHRASE
Put this file in ``/usr/local/sbin/backup.sh``. Run ``crontab -e`` and
paste your configuration into the file that opens. If you're unfamiliar
with Cron, here is a good `HowTo
<https://help.ubuntu.com/community/CronHowto>`__. If the folder being
backed up is a folder to be modified permanently during the work day,
we can set incremental backups every 5 minutes from 8 AM to 9 PM Monday
through Friday by pasting the following line into crontab:
So let's say you put this file in ``/usr/local/sbin/backup.sh.`` Next
you want to run ``crontab -e`` and paste your configuration in the file
that opens. If you're unfamiliar with Cron, here is a good `How
To <https://help.ubuntu.com/community/CronHowto>`__. The folder I'm
backing up is a folder I modify permanently during my workday, so I want
incremental backups every 5mn from 8AM to 9PM monday to friday. Here is
the line I will paste in my crontab:
.. code:: sh
.. code:: cron
*/5 8-20 * * 1-5 /usr/local/sbin/backup.sh
Adding or removing files from the folder being backed up will result in
incremental backups in the bucket.
Now I can try and add / remove files from the folder I'm backing up, and
I will see incremental backups in my bucket.

View File

@ -1,263 +0,0 @@
Metadata Search Documentation
=============================
Description
-----------
This feature enables metadata search to be performed on the metadata of objects
stored in Zenko.
Requirements
------------
* MongoDB
Design
------
The Metadata Search feature expands on the existing :code:`GET Bucket` S3 API by
enabling users to conduct metadata searches by adding the custom Zenko query
string parameter, :code:`search`. The :code:`search` parameter is structured as a pseudo
SQL WHERE clause, and supports basic SQL operators. For example:
:code:`"A=1 AND B=2 OR C=3"` (complex queries can be built using nesting
operators, :code:`(` and :code:`)`).
The search process is as follows:
* Zenko receives a :code:`GET` request.
.. code::
# regular getBucket request
GET /bucketname HTTP/1.1
Host: 127.0.0.1:8000
Date: Wed, 18 Oct 2018 17:50:00 GMT
Authorization: authorization string
# getBucket versions request
GET /bucketname?versions HTTP/1.1
Host: 127.0.0.1:8000
Date: Wed, 18 Oct 2018 17:50:00 GMT
Authorization: authorization string
# search getBucket request
GET /bucketname?search=key%3Dsearch-item HTTP/1.1
Host: 127.0.0.1:8000
Date: Wed, 18 Oct 2018 17:50:00 GMT
Authorization: authorization string
* If the request does *not* contain the :code:`search` query parameter, Zenko performs
a normal bucket listing and returns an XML result containing the list of
objects.
* If the request *does* contain the :code:`search` query parameter, Zenko parses and
validates the search string.
- If the search string is invalid, Zenko returns an :code:`InvalidArgument` error.
.. code::
<?xml version=\"1.0\" encoding=\"UTF-8\"?>
<Error>
<Code>InvalidArgument</Code>
<Message>Invalid sql where clause sent as search query</Message>
<Resource></Resource>
<RequestId>d1d6afc64345a8e1198e</RequestId>
</Error>
- If the search string is valid, Zenko parses it and generates an abstract
syntax tree (AST). The AST is then passed to the MongoDB backend to be
used as the query filter for retrieving objects from a bucket that
satisfies the requested search conditions. Zenko parses the filtered
results and returns them as the response.
Metadata search results have the same structure as a :code:`GET Bucket` response:
.. code:: xml
<?xml version="1.0" encoding="UTF-8"?>
<ListBucketResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/">
<Name>bucketname</Name>
<Prefix/>
<Marker/>
<MaxKeys>1000</MaxKeys>
<IsTruncated>false</IsTruncated>
<Contents>
<Key>objectKey</Key>
<LastModified>2018-04-19T18:31:49.426Z</LastModified>
<ETag>&quot;d41d8cd98f00b204e9800998ecf8427e&quot;</ETag>
<Size>0</Size>
<Owner>
<ID>79a59df900b949e55d96a1e698fbacedfd6e09d98eacf8f8d5218e7cd47ef2be</ID>
<DisplayName>Bart</DisplayName>
</Owner>
<StorageClass>STANDARD</StorageClass>
</Contents>
<Contents>
...
</Contents>
</ListBucketResult>
Performing Metadata Searches with Zenko
---------------------------------------
You can perform metadata searches by:
+ Using the :code:`search_bucket` tool in the
`Scality/S3 <https://github.com/scality/S3>`_ GitHub repository.
+ Creating a signed HTTP request to Zenko in your preferred programming
language.
Using the S3 Tool
+++++++++++++++++
After cloning the `Scality/S3 <https://github.com/scality/S3>`_ GitHub repository
and installing the necessary dependencies, run the following command in the S3
projects root directory to access the search tool:
.. code::
node bin/search_bucket
This generates the following output:
.. code::
Usage: search_bucket [options]
Options:
-V, --version output the version number
-a, --access-key <accessKey> Access key id
-k, --secret-key <secretKey> Secret access key
-b, --bucket <bucket> Name of the bucket
-q, --query <query> Search query
-h, --host <host> Host of the server
-p, --port <port> Port of the server
-s --ssl
-v, --verbose
-h, --help output usage information
In the following examples, Zenko Server is accessible on endpoint
:code:`http://127.0.0.1:8000` and contains the bucket :code:`zenkobucket`.
.. code::
# search for objects with metadata "blue"
node bin/search_bucket -a accessKey1 -k verySecretKey1 -b zenkobucket \
-q "x-amz-meta-color=blue" -h 127.0.0.1 -p 8000
# search for objects tagged with "type=color"
node bin/search_bucket -a accessKey1 -k verySecretKey1 -b zenkobucket \
-q "tags.type=color" -h 127.0.0.1 -p 8000
Coding Examples
+++++++++++++++
Search requests can be also performed by making HTTP requests authenticated
with one of the AWS Signature schemes: version 2 or version 4. \
For more about authentication scheme, see:
* https://docs.aws.amazon.com/general/latest/gr/signature-version-2.html
* http://docs.aws.amazon.com/general/latest/gr/sigv4_signing.html
* http://docs.aws.amazon.com/general/latest/gr/sigv4-signed-request-examples.html
You can also view examples for making requests with Auth V4 in various
languages `here <../../../examples>`__.
Specifying Metadata Fields
~~~~~~~~~~~~~~~~~~~~~~~~~~
To search system metadata headers:
.. code::
{system-metadata-key}{supported SQL op}{search value}
# example
key = blueObject
size > 0
key LIKE "blue.*"
To search custom user metadata:
.. code::
# metadata must be prefixed with "x-amz-meta-"
x-amz-meta-{user-metadata-key}{supported SQL op}{search value}
# example
x-amz-meta-color = blue
x-amz-meta-color != red
x-amz-meta-color LIKE "b.*"
To search tags:
.. code::
# tag searches must be prefixed with "tags."
tags.{tag-key}{supported SQL op}{search value}
# example
tags.type = color
Examples queries:
.. code::
# searching for objects with custom metadata "color"=blue" and are tagged
# "type"="color"
tags.type="color" AND x-amz-meta-color="blue"
# searching for objects with the object key containing the substring "blue"
# or (custom metadata "color"=blue" and are tagged "type"="color")
key LIKE '.*blue.*' OR (x-amz-meta-color="blue" AND tags.type="color")
Differences from SQL
++++++++++++++++++++
Zenko metadata search queries are similar to SQL-query :code:`WHERE` clauses, but
differ in that:
* They follow the :code:`PCRE` format
* They do not require values with hyphens to be enclosed in
backticks, :code:``(`)``
.. code::
# SQL query
`x-amz-meta-search-item` = `ice-cream-cone`
# MD Search query
x-amz-meta-search-item = ice-cream-cone
* Search queries do not support all SQL operators.
.. code::
# Supported SQL operators:
=, <, >, <=, >=, !=, AND, OR, LIKE, <>
# Unsupported SQL operators:
NOT, BETWEEN, IN, IS, +, -, %, ^, /, *, !
Using Regular Expressions in Metadata Search
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Regular expressions in Zenko metadata search differ from SQL in the following
ways:
+ Wildcards are represented with :code:`.*` instead of :code:`%`.
+ Regex patterns must be wrapped in quotes. Failure to do this can lead to
misinterpretation of patterns.
+ As with :code:`PCRE`, regular expressions can be entered in either the
:code:`/pattern/` syntax or as the pattern itself if regex options are
not required.
Example regular expressions:
.. code::
# search for strings containing word substring "helloworld"
".*helloworld.*"
"/.*helloworld.*/"
"/.*helloworld.*/i"

Some files were not shown because too many files have changed in this diff Show More