e77f8e311c
etcdserver: parse context error for better message
2014-10-27 10:59:15 -07:00
009b737cef
*: better logging
2014-10-26 08:13:03 -07:00
719c57a29d
proxy: retrieve ClientURLs from cluster
...
This is a simple solution to having the proxy keep up to date with the
state of the cluster. Basically, it uses the cluster configuration
provided at start up (i.e. with `-initial-cluster-state`) to determine
where to reach peer(s) in the cluster, and then it will periodically hit
the `/members` endpoint of those peer(s) (using the same mechanism that
`-cluster-state=existing` does to initialise) to update the set of valid
client URLs to proxy to.
This does not address discovery (#1376 ), and it would probably be better
to update the set of proxyURLs dynamically whenever we fetch the new
state of the cluster; but it needs a bit more thinking to have this done
in a clean way with the proxy interface.
Example in Procfile works again.
2014-10-24 15:54:12 -07:00
49a68adcf1
Merge pull request #1386 from unihorn/184
...
etcdserver: update member attribute when apply request
2014-10-24 12:46:11 -07:00
ea0bff80c0
etcdserver: update member attribute when apply request
2014-10-24 12:43:53 -07:00
08593bcdf6
etcdserver: support newly-join member bootstrap
2014-10-24 12:38:44 -07:00
4d80f01201
etcdserver: Cluster.IsIDremoved -> Cluster.IsIDRemoved
2014-10-23 14:29:58 -07:00
8eee8c260e
etcdserver: rebase on master and code clean
2014-10-23 13:58:55 -07:00
f8b8bdeb17
etcdserver: use path.Join for member key in cluster
2014-10-23 13:27:54 -07:00
3d243baacd
etcdserver: generate id when new cluster
2014-10-23 13:27:54 -07:00
67412e07f8
etcdserver: MemberFromName -> MemberByName
2014-10-23 13:27:54 -07:00
89572b5fd7
etcdserver: refactor cluster and clusterStore
...
Integrate clusterStore into cluster, and let cluster become the source of
cluster info.
2014-10-23 13:27:54 -07:00
ab90369f9e
etcdserver: use hex for cluster and machine id
...
Continue using hex everywhere. Including here.
TODO: cleanup the printing of the structs which currently have decimal
to/from:
`{Type:MsgAppResp To:9973738105406047488 From:17050684879817348455 T...`
2014-10-22 16:24:50 -07:00
e42d65da12
etcdserver: Check the initial cluster settings after checking if the WAL exists
2014-10-22 18:16:43 -04:00
13656eb4e7
Merge pull request #1340 from barakmich/better_ids2
...
etcdserver: Calculate IDs based on PeerURLs and --initial-cluster-name
2014-10-22 14:49:49 -04:00
89b032cd69
etcdserver: Member.storeKey -> memberStoreKey
2014-10-22 11:09:36 -07:00
7498234e40
etcdserver: record removed member to check incoming message
2014-10-22 11:09:35 -07:00
502a3c2460
Refactor Cluster to hold and add members.
2014-10-22 13:52:42 -04:00
1ca7c031ff
first round of comments
...
Conflicts:
etcdserver/config.go
etcdserver/config_test.go
etcdserver/server.go
main.go
2014-10-22 13:49:54 -04:00
456d1ebcae
etcdserver: Calculate IDs for nodes solely on PeerURLs
...
Removes the notion of name being anything more than advisory or
command-line grouping, and adds checks for bootstrapping the command
line. IDs are consistent if the URLs are consistent.
2014-10-22 13:49:54 -04:00
e2b6a4fc4c
etcdserver: const XXXDir -> StoreXXXPrefix
...
and code clean
2014-10-21 16:10:49 -07:00
2ff3cac653
etcdserver/etcdhttp: store location adjustment
...
Detailed adjustment:
/_etcd/machines/* -> /0/members/*
/* -> /1/*
And it keeps key path returned to user the same as before.
2014-10-21 16:10:19 -07:00
894e678ad6
etcdserver: checking clusterID
2014-10-21 11:05:24 -07:00
a44849deec
Merge pull request #1286 from coreos/clusterid
...
*: generate clusterid
2014-10-20 19:07:03 -07:00
0fd28169c8
etcdserver: use id,cid
2014-10-20 16:35:41 -07:00
dc68dc9ebd
etcdserver: add a todo for clusterid generation
2014-10-20 16:26:31 -07:00
92230cee63
Merge pull request #1333 from unihorn/172
...
etcdserver/raft: remove msgDenied, removedNodes, shouldStop
2014-10-20 16:04:09 -07:00
e200d2a8e2
etcdserver/raft: remove msgDenied, removedNodes, shouldStop
...
The future plan is to do all these in etcdserver level.
2014-10-20 15:13:18 -07:00
ea6bcacfe4
*: generate clusterid
2014-10-20 15:00:54 -07:00
f7a0d5387b
etcdserver: stop server gracefully
2014-10-20 14:43:21 -07:00
058537f34a
main: add basic functional test
2014-10-20 14:43:21 -07:00
dcaa7f0a37
Merge pull request #1328 from unihorn/169
...
skip initial-cluster check when reboot
2014-10-20 13:06:24 -07:00
17382ec905
etcdserver: skip initial-cluster check when reboot
...
If etcd is provided with data-dir that has data, it will not use
initial-cluster, and initial-cluster could be set to be empty.
2014-10-20 12:41:07 -07:00
7a4d42166b
*: add license header to all source files
2014-10-17 15:41:22 -07:00
c28907ba95
etcdserver: fix race and improve stats interfaces
2014-10-16 23:05:48 -07:00
c30b82b596
etcdserver: fix data race in retrieving self stats
2014-10-16 23:05:48 -07:00
97ae531eda
etcdserver: split out storestats and serverstats
2014-10-16 10:43:49 -07:00
fedb67a71a
etcdserver: fix TODO and change to base 16
2014-10-16 10:43:49 -07:00
8168fed825
etcdserver: add ServerStats and LeaderStats
...
This adds the remaining two stats endpoints: `/v2/stats/self`, for
various statistics on the EtcdServer, and `/v2/stats/leader`, for
statistics on a leader's followers.
By and large most of the stats code is copied across from 0.4.x, updated
where necessary to integrate with the new decoupling of raft from
transport.
This does not satisfactorily resolve the question of name vs ID. In the
old world, names were unique in the cluster and transmitted over the
wire, so they could be used safely in all statistics. In the new world,
a given EtcdServer only knows its own name, and it is instead IDs that
are communicated among the cluster members. Hence in most places here we
simply substitute a string-encoded ID in place of name, and only where
possible do we retain the actual given name of the EtcdServer.
2014-10-16 10:43:44 -07:00
0a8721a708
etcdserver: expose store statistics
2014-10-16 10:43:23 -07:00
8cd6030a1d
etcdserver: add checking when apply conf change
2014-10-16 09:49:26 -07:00
a8a1d4fd93
Merge pull request #1299 from unihorn/162
...
etcdhttp: add PUT and DELETE on /v2/admin/members/
2014-10-15 10:52:11 -07:00
828accf07b
etcdserver: better logging
2014-10-15 13:07:34 +08:00
f98fbbfc14
*: proto refactoring
2014-10-14 21:07:23 +08:00
4183b69e12
*: move from third_party to Godep
2014-10-14 00:37:52 -07:00
31264e7eb5
etcdhttp: add PUT and DELETE on /v2/admin/members/
2014-10-13 22:36:06 -07:00
6fc0b1977b
etcdserver: Id -> ID
2014-10-14 11:30:49 +08:00
b53b74733a
etcdserver: add s.apply
2014-10-14 11:30:49 +08:00
30c7a7f2dd
etcdserver: add shouldDiscover
2014-10-14 11:29:00 +08:00
a85ec90d68
etcdserver: clean NewServer
2014-10-14 11:29:00 +08:00
f693c6ddf2
etcdserver: apply bootstrap conf change
2014-10-13 11:22:23 -07:00
0319b033ea
etcdserver/raft: set context for bootstrap addnode entries
2014-10-13 11:22:23 -07:00
3516cc3ee5
*: change all ID to uint64
2014-10-12 08:38:48 +08:00
f16a272898
Merge pull request #1234 from unihorn/152
...
etcdserver: save PeerURLs and Others into two different keys
2014-10-10 12:21:32 -07:00
447caf1afc
etcdserver/wal: record info at the head of WAL file
2014-10-10 11:57:09 -07:00
77271b0663
etcdserver: split Member into RaftAttributes and Attributes
...
The split helps to save them in different key paths, because they have
distinct life cycle on update.
2014-10-09 20:28:00 -07:00
8bbbaa88b2
*: raft related int64 -> uint64
2014-10-09 14:29:21 +08:00
682008724d
etcdserver: add discovery warning message
2014-10-08 16:09:12 -07:00
8e8719f6ac
etcdserver: Create, Delete -> Add, Remove
2014-10-07 16:21:58 -07:00
cdea98d434
etcdserver: skip bootstrap addNode entry
...
This is a hack to make etcd work normally.
2014-10-07 11:11:31 -07:00
6a62621695
etcdserver: update panic info
2014-10-07 11:04:50 -07:00
3859297225
etcdserver: check id match
2014-10-07 11:04:50 -07:00
d051af4d3d
etcdserver: apply config change on cluster store
2014-10-07 11:04:50 -07:00
5587e0d73f
raft: compact takes index and nodes parameters
...
Before this commit, compact always compact log at current appliedindex of raft.
This prevents us from doing non-blocking snapshot since we have to make snapshot
and compact atomically. To prepare for non-blocking snapshot, this commit make
compact supports index and nodes parameters. After completing snapshot, the applier
should call compact with the snapshot index and the nodes at snapshot index to do
a compaction at snapsohot index.
2014-10-07 16:03:11 +08:00
39e0a0cd0a
Merge pull request #1249 from barakmich/sanity_check
...
Split config into a separate file and add sanity check and test
2014-10-06 20:19:23 -04:00
3ca3c9ad4c
Merge pull request #1245 from unihorn/155
...
main/raft: write addNode ConfChange entries in log when start raft
2014-10-06 17:00:28 -07:00
120b088723
Split config into a separate file and add sanity check and test
2014-10-06 19:42:13 -04:00
45ebfb4217
raft: refine initial entries logic in StartNode
2014-10-06 16:06:01 -07:00
c15c3eab4c
etcdserver: move int64Slice into pkg/types/
2014-10-06 15:12:02 -07:00
e2d8037ded
main: use initial-cluster and initial-cluster-state flags
...
In preperation for adding the ability to join a machine to an existing
cluster force the user to specify whether they expect this to me a new
cluster or an active one.
The error for not specifying the initial-cluster-state is:
```
etcd: initial cluster state unset and no wal found
```
2014-10-06 14:59:25 -07:00
314d425718
main/raft: write addNode ConfChange entries in log when start raft
2014-10-06 14:33:12 -07:00
9e3d045b2b
*:discovery hook up
2014-10-06 10:12:42 -07:00
1308c3e809
etcdserver: unexport EtcdServer fields where possible
2014-10-03 13:34:56 -07:00
e5f5fcff48
etcdserver: collapse dir configuration to DataDir
2014-10-03 13:34:49 -07:00
af6b29f291
main: move server configuration to etcdserver package
2014-10-03 13:34:49 -07:00
e4a6c9651a
raft: add removed
...
The usage of removed:
1. tell removed node about its removal explicitly using msgDenied
2. prevent removed node disrupt cluster progress by launching leader election
It is set when apply node removal, or receive msgDenied.
2014-10-01 14:57:38 -07:00
83137f9eba
pkg/types: introduce a URLs type
...
Cleanup the usage of URLs into its own type so we don't have to use a
FlagValue everywhere we have a list of URLs.
2014-10-01 14:41:01 -07:00
073eb7677d
etcdserver: move grep member logic into publish func
2014-10-01 11:50:47 -07:00
cbc84bc70e
etcdserver: minimize input info
2014-10-01 11:47:17 -07:00
a40a270e19
etcdserver: publish self info when start
2014-10-01 11:47:17 -07:00
34547229a6
etcdserver: fix left Id -> ID
2014-09-29 14:14:45 -07:00
0e8345aa73
Merge pull request #1143 from unihorn/136
...
*: Id -> ID for protobuf types
2014-09-29 13:58:02 -07:00
89077167c3
etcdserver: add publish func
2014-09-29 10:13:06 -07:00
1d5d2e3726
*: Id -> ID for protobuf types
...
We use ID instead of Id in this project based on golang conventions.
2014-09-26 11:49:30 -07:00
5e3fd6ee3f
etcdserver: introduce the cluster and member
...
This introduces two new concepts: the cluster and the member.
Members are logical etcd instances that have a name, raft ID, and a list
of peer and client addresses.
A cluster is made up of a list of members.
2014-09-25 18:04:10 -07:00
9b3478218e
etcdserver: add RaftIndex and RaftTerm
2014-09-25 16:46:24 -07:00
c28fd92d10
etcdserver: extend storeRecorder to test parameters
2014-09-24 14:33:51 -07:00
172a32e5e3
etcdserver: correct timeout and streaming handling
...
This reintroduces the 'stream' parameter to support long-lived watch
sessions. These sessions respect a server timeout (set to 5 minutes by
default).
2014-09-24 11:27:36 -07:00
bc7b0108dc
raft: ConfigChange -> ConfChange
2014-09-23 12:02:44 -07:00
d92931853e
raft: Config -> ConfigChange
...
Configure -> ProposeConfigChange
AddNode, RemoveNode -> ApplyConfigChange
2014-09-22 23:39:53 -07:00
ec8f493fde
raft: refine comments for Configure
2014-09-22 15:44:47 -07:00
dc36ae7058
raft: use pb.Config instead of []byte for Configure
2014-09-22 15:44:47 -07:00
4203569da2
etcdserver: use ConfigType in protobuf
2014-09-22 15:44:46 -07:00
b82d70871f
raft: use EntryType in protobuf
2014-09-22 15:44:46 -07:00
abdb2cad15
etcdserver: Config.Id -> Config.ID
2014-09-22 15:44:46 -07:00
aaffb9eb78
etcdserver: add AddNode, RemoveNode
...
AddNode and RemoveNode is used to propose config change to the cluster.
If succeeds, it will add/remove node from the cluster.
2014-09-22 15:44:46 -07:00
7d1126fb35
etcdserver: init time.Time only if Expiration > 0
...
golang's concept of "zero" time does not align with the zero value of
a unix timestamp. Initializing time.Time with a unix timestamp of 0
makes time.Time.IsZero fail. Solve this by initializing time.Time only
if we care about the time.
2014-09-22 15:31:31 -07:00
e08df4c8d2
store: fix recursive/dir arguments
2014-09-22 14:44:01 -07:00
91003cb994
*: s/prevExists/prevExist/
2014-09-18 21:25:08 -07:00
a42d52482c
etcdserver: use simple PRNG for GenID
2014-09-17 15:29:14 -07:00
ab61a8aa9a
*: init for on disk snap support
2014-09-17 13:56:12 -07:00
023dc7cba2
etcdserver: add SYNC request
2014-09-16 13:42:03 -07:00
8a5ab2ec06
etcdserver: introduce Server interface
...
This changes etcdserver.Server to an interface, with the former Server
(now "EtcdServer") becoming the canonical/production implementation.
This will facilitate better testing of the http server et al with mock
implementations of the interface.
It also more clearly defines the boundary for users of the Server.
2014-09-15 15:11:01 -07:00
6cd4434ff3
server: add unit tests
...
Make test coverage >= 90%
2014-09-15 13:16:48 -07:00
f7444ff300
store: convert Watch to interface
2014-09-09 11:17:53 -07:00
d844377ca6
*: vendor context
2014-09-03 20:06:16 -07:00
2ba57ee75d
*: remove unused pkgs
2014-09-03 09:20:20 -07:00
Yicheng Qin