Fix pool create/modify --block_size validation

A bunch of monitor fixes

- Add noout flag for OSDs (/vitastor/config/osd/xx)
- Fix "effective" size of degraded PGs (and thus "used space") calculation in monitor
- Fix monitor not clearing PGs of deleted pools
- Fix incorrect PG generation with hosts with 0 OSDs
- Fix monitor crashing during primary OSD recheck when pool has no PGs
- Fix monitor crashing when node_placement included non-existing OSDs
- Fix possible data movement after removing OSDs reweighted to 0
- Remove extra empty keys from pool configurations created by vitastor-cli create-pool
- Fix 32-bit build

.gitea/workflows/test.yml

@@ -64,6 +64,13 @@ jobs:
     # leak sanitizer sometimes crashes
     - run: cd /root/vitastor/build && ASAN_OPTIONS=detect_leaks=0 make -j16 test
 
+  npm_lint:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - run: cd /root/vitastor/mon && npm run lint
+
   test_add_osd:
     runs-on: ubuntu-latest
     needs: build

CMakeLists.txt

@@ -2,6 +2,6 @@ cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
-set(VERSION "1.6.0")
+set(VERSION "1.6.1")
 
 add_subdirectory(src)

csi/Makefile

@@ -1,4 +1,4 @@
-VERSION ?= v1.6.0
+VERSION ?= v1.6.1
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v1.6.0
+          image: vitalif/vitastor-csi:v1.6.1
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -121,7 +121,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v1.6.0
+          image: vitalif/vitastor-csi:v1.6.1
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "1.6.0"
+    vitastorCSIDriverVersion = "1.6.1"
 )
 
 // Config struct fills the parameters of request or user input

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (1.6.0-1) unstable; urgency=medium
+vitastor (1.6.1-1) unstable; urgency=medium
 
   * Bugfixes
 

debian/vitastor.Dockerfile

@@ -37,8 +37,8 @@ RUN set -e -x; \
     mkdir -p /root/packages/vitastor-$REL; \
     rm -rf /root/packages/vitastor-$REL/*; \
     cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-1.6.0; \
-    cd vitastor-1.6.0; \
+    cp -r /root/vitastor vitastor-1.6.1; \
+    cd vitastor-1.6.1; \
     ln -s /root/fio-build/fio-*/ ./fio; \
     FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
@@ -51,8 +51,8 @@ RUN set -e -x; \
     rm -rf a b; \
     echo "dep:fio=$FIO" > debian/fio_version; \
     cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_1.6.0.orig.tar.xz vitastor-1.6.0; \
-    cd vitastor-1.6.0; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_1.6.1.orig.tar.xz vitastor-1.6.1; \
+    cd vitastor-1.6.1; \
     V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     DEBFULLNAME="Vitaliy Filippov <vitalif@yourcmc.ru>" dch -D $REL -v "$V""$REL" "Rebuild for $REL"; \
     DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage --jobs=auto -sa; \

docs/config/pool.en.md

@@ -86,7 +86,11 @@ Parent node reference is required for intermediate tree nodes.
 Separate OSD settings are set in etc keys `/vitastor/config/osd/<number>`
 in JSON format `{"<key>":<value>}`.
 
-As of now, two settings are supported:
+As of now, the following settings are supported:
+
+- [reweight](#reweight)
+- [tags](#tags)
+- [noout](#noout)
 
 ## reweight
 
@@ -109,6 +113,14 @@ subsets and then use a specific subset for pool instead of all OSDs.
 For example you can mark SSD OSDs with tag "ssd" and HDD OSDs with "hdd" and
 such tags will work as device classes.
 
+## noout
+
+- Type: boolean
+- Default: false
+
+If set to true, [osd_out_time](monitor.en.md#osd_out_time) is ignored for this
+OSD and it's never removed from data distribution by the monitor.
+
 # Pool parameters
 
 ## name

docs/config/pool.ru.md

@@ -85,10 +85,11 @@
 Настройки отдельных OSD задаются в ключах etcd `/vitastor/config/osd/<number>`
 в JSON-формате `{"<key>":<value>}`.
 
-На данный момент поддерживаются две настройки:
+На данный момент поддерживаются следующие настройки:
 
 - [reweight](#reweight)
 - [tags](#tags)
+- [noout](#noout)
 
 ## reweight
 
@@ -112,6 +113,14 @@
 всех. Можно, например, пометить SSD OSD тегом "ssd", а HDD тегом "hdd", в
 этом смысле теги работают аналогично классам устройств.
 
+## noout
+
+- Тип: булево (да/нет)
+- Значение по умолчанию: false
+
+Если установлено в true, то [osd_out_time](monitor.ru.md#osd_out_time) для этого
+OSD игнорируется и OSD не удаляется из распределения данных монитором.
+
 # Параметры
 
 ## name

mon/.eslintrc.js

@@ -0,0 +1,49 @@
+module.exports = {
+    "env": {
+        "es6": true,
+        "node": true
+    },
+    "extends": [
+        "eslint:recommended",
+        "plugin:node/recommended"
+    ],
+    "parserOptions": {
+        "ecmaVersion": 2020
+    },
+    "plugins": [
+    ],
+    "rules": {
+        "indent": [
+            "error",
+            4
+        ],
+        "brace-style": [
+            "error",
+            "allman",
+            { "allowSingleLine": true }
+        ],
+        "linebreak-style": [
+            "error",
+            "unix"
+        ],
+        "semi": [
+            "error",
+            "always"
+        ],
+        "no-useless-escape": [
+            "off"
+        ],
+        "no-control-regex": [
+            "off"
+        ],
+        "no-empty": [
+            "off"
+        ],
+        "no-process-exit": [
+            "off"
+        ],
+        "node/shebang": [
+            "off"
+        ]
+    }
+};

mon/PGUtil.js

@@ -97,7 +97,6 @@ function scale_pg_history(prev_pg_history, prev_pgs, new_pgs)
 
 function scale_pg_count(prev_pgs, new_pg_count)
 {
-    const old_pg_count = prev_pgs.length;
     // Just for the lp_solve optimizer - pick a "previous" PG for each "new" one
     if (prev_pgs.length < new_pg_count)
     {

mon/dsl_pgs.js

@@ -152,6 +152,7 @@ function parse_pg_dsl(text)
             else if (rule[2] === '(')
             {
                 rule[2] = [];
+                // eslint-disable-next-line no-constant-condition
                 while (true)
                 {
                     if (i > tokens.length-1)

mon/lp-optimizer.js

@@ -77,7 +77,7 @@ async function optimize_initial({ osd_weights, combinator, pg_count, pg_size = 3
     {
         if (osd !== NO_OSD)
         {
-            let osd_pg_count = (osd_weights[osd]||0)/total_weight*pg_effsize*pg_count;
+            let osd_pg_count = osd_weights[osd]/total_weight*pg_effsize*pg_count;
             lp += pg_per_osd[osd].join(' + ')+' <= '+osd_pg_count+';\n';
         }
     }
@@ -215,7 +215,7 @@ function calc_intersect_weights(old_pg_size, pg_size, pg_count, prev_weights, al
                 {
                     const intersect_count = ordered
                         ? pg.reduce((a, osd, i) => a + (prev_hash[osd] == 1+i ? 1 : 0), 0)
-                        : pg.reduce((a, osd, i) => a + (prev_hash[osd] ? 1 : 0), 0);
+                        : pg.reduce((a, osd) => a + (prev_hash[osd] ? 1 : 0), 0);
                     if (max_int < intersect_count)
                     {
                         max_int = intersect_count;
@@ -299,7 +299,7 @@ async function optimize_change({ prev_pgs: prev_int_pgs, osd_weights, combinator
             )).join(' + ');
             const rm_osd_pg_count = (prev_pg_per_osd[osd]||[])
                 .reduce((a, [ old_pg_name, space ]) => (a + (all_pgs_hash[old_pg_name] ? space : 0)), 0);
-            const osd_pg_count = (osd_weights[osd]||0)*pg_effsize/total_weight*pg_count - rm_osd_pg_count;
+            const osd_pg_count = osd_weights[osd]*pg_effsize/total_weight*pg_count - rm_osd_pg_count;
             lp += osd_sum + ' <= ' + osd_pg_count + ';\n';
         }
     }

mon/mon.js

@@ -37,7 +37,6 @@ const etcd_allow = new RegExp('^'+[
     'pg/state/[1-9]\\d*/[1-9]\\d*',
     'pg/stats/[1-9]\\d*/[1-9]\\d*',
     'pg/history/[1-9]\\d*/[1-9]\\d*',
-    'pool/stats/[1-9]\\d*',
     'history/last_clean_pgs',
     'inode/stats/[1-9]\\d*/\\d+',
     'pool/stats/[1-9]\\d*',
@@ -215,7 +214,7 @@ const etcd_tree = {
         }, */
         pools: {},
         osd: {
-            /* <id>: { reweight?: 1, tags?: [ 'nvme', ... ] }, ... */
+            /* <id>: { reweight?: 1, tags?: [ 'nvme', ... ], noout?: true }, ... */
         },
         /* pgs: {
             hash: string,
@@ -585,7 +584,7 @@ class Mon
                 now = Date.now();
             }
             tried[base] = now;
-            const ok = await new Promise((ok, no) =>
+            const ok = await new Promise(ok =>
             {
                 const timer_id = setTimeout(() =>
                 {
@@ -747,6 +746,7 @@ class Mon
         this.save_last_clean_running = true;
         // last_clean_pgs is used to avoid extra data move when observing a series of changes in the cluster
         const new_clean_pgs = { items: {} };
+        // eslint-disable-next-line indent
     next_pool:
         for (const pool_id in this.state.config.pools)
         {
@@ -829,6 +829,7 @@ class Mon
     async become_master()
     {
         const state = { ...this.get_mon_state(), id: ''+this.etcd_lease_id };
+        // eslint-disable-next-line no-constant-condition
         while (1)
         {
             const res = await this.etcd_call('/kv/txn', {
@@ -873,32 +874,19 @@ class Mon
         levels.osd = levels.osd || 101;
         const tree = {};
         let up_osds = {};
-        for (const node_id in this.state.config.node_placement||{})
-        {
-            const node_cfg = this.state.config.node_placement[node_id];
-            if (/^\d+$/.exec(node_id))
-            {
-                node_cfg.level = 'osd';
-            }
-            if (!node_id || !node_cfg.level || !levels[node_cfg.level])
-            {
-                // All nodes must have non-empty IDs and valid levels
-                continue;
-            }
-            tree[node_id] = { id: node_id, level: node_cfg.level, parent: node_cfg.parent, children: [] };
-        }
         // This requires monitor system time to be in sync with OSD system times (at least to some extent)
         const down_time = Date.now()/1000 - this.config.osd_out_time;
         for (const osd_num of this.all_osds().sort((a, b) => a - b))
         {
             const stat = this.state.osd.stats[osd_num];
-            if (stat && stat.size && (this.state.osd.state[osd_num] || Number(stat.time) >= down_time))
+            const osd_cfg = this.state.config.osd[osd_num];
+            let reweight = osd_cfg == null ? 1 : Number(osd_cfg.reweight);
+            if (reweight < 0 || isNaN(reweight))
+                reweight = 1;
+            if (stat && stat.size && reweight && (this.state.osd.state[osd_num] || Number(stat.time) >= down_time ||
+                osd_cfg && osd_cfg.noout))
             {
                 // Numeric IDs are reserved for OSDs
-                const osd_cfg = this.state.config.osd[osd_num];
-                let reweight = osd_cfg == null ? 1 : Number(osd_cfg.reweight);
-                if (reweight < 0 || isNaN(reweight))
-                    reweight = 1;
                 if (this.state.osd.state[osd_num] && reweight > 0)
                 {
                     // React to down OSDs immediately
@@ -926,6 +914,29 @@ class Mon
                 }
             }
         }
+        for (const node_id in this.state.config.node_placement||{})
+        {
+            const node_cfg = this.state.config.node_placement[node_id];
+            if (/^\d+$/.exec(node_id))
+            {
+                node_cfg.level = 'osd';
+            }
+            if (!node_id || !node_cfg.level || !levels[node_cfg.level] ||
+                node_cfg.level === 'osd' && !tree[node_id])
+            {
+                // All nodes must have non-empty IDs and valid levels
+                // OSDs have to actually exist
+                continue;
+            }
+            tree[node_id] = tree[node_id] || {};
+            tree[node_id].id = node_id;
+            tree[node_id].level = node_cfg.level;
+            tree[node_id].parent = node_cfg.parent;
+            if (node_cfg.level !== 'osd')
+            {
+                tree[node_id].children = [];
+            }
+        }
         return { up_osds, levels, osd_tree: tree };
     }
 
@@ -955,6 +966,25 @@ class Mon
             const parent = parent_level && parent_level < node_level ? node_cfg.parent : '';
             tree[parent].children.push(tree[node_id]);
         }
+        // Delete empty nodes
+        let deleted = 0;
+        do
+        {
+            deleted = 0;
+            for (const node_id in tree)
+            {
+                if (tree[node_id].level !== 'osd' && (!tree[node_id].children || !tree[node_id].children.length))
+                {
+                    const parent = tree[node_id].parent;
+                    if (parent)
+                    {
+                        tree[parent].children = tree[parent].children.filter(c => c != tree[node_id]);
+                    }
+                    deleted++;
+                    delete tree[node_id];
+                }
+            }
+        } while (deleted > 0);
         return tree;
     }
 
@@ -991,7 +1021,7 @@ class Mon
                 const key = b64(this.etcd_prefix+'/osd/state/'+osd_num);
                 checks.push({ key, target: 'MOD', result: 'LESS', mod_revision: ''+this.etcd_watch_revision });
             }
-            const res = await this.etcd_call('/kv/txn', {
+            await this.etcd_call('/kv/txn', {
                 compare: [
                     { key: b64(this.etcd_prefix+'/mon/master'), target: 'LEASE', lease: ''+this.etcd_lease_id },
                     { key: b64(this.etcd_prefix+'/config/pgs'), target: 'MOD', mod_revision: ''+this.etcd_watch_revision, result: 'LESS' },
@@ -1315,7 +1345,7 @@ class Mon
             for (const k in rules)
             {
                 if (!levels[k] || typeof rules[k] !== 'string' &&
-                    (!rules[k] instanceof Array ||
+                    (!(rules[k] instanceof Array) ||
                     rules[k].filter(s => typeof s !== 'string' && typeof s !== 'number').length > 0))
                 {
                     if (warn)
@@ -1429,7 +1459,15 @@ class Mon
         }
         console.log(`Pool ${pool_id} (${pool_cfg.name || 'unnamed'}):`);
         LPOptimizer.print_change_stats(optimize_result);
-        const pg_effsize = Math.min(pool_cfg.pg_size, Object.keys(pool_tree).length);
+        let pg_effsize = pool_cfg.pg_size;
+        for (const pg of optimize_result.int_pgs)
+        {
+            const this_pg_size = pg.filter(osd => osd != LPOptimizer.NO_OSD).length;
+            if (this_pg_size && this_pg_size < pg_effsize)
+            {
+                pg_effsize = this_pg_size;
+            }
+        }
         return {
             pool_id,
             pgs: optimize_result.int_pgs,
@@ -1511,11 +1549,14 @@ class Mon
                 {
                     continue;
                 }
-                const replicated = pool_cfg.scheme === 'replicated';
                 const aff_osds = this.get_affinity_osds(pool_cfg, up_osds, osd_tree);
                 this.reset_rng();
                 for (let pg_num = 1; pg_num <= pool_cfg.pg_count; pg_num++)
                 {
+                    if (!this.state.config.pgs.items[pool_id])
+                    {
+                        continue;
+                    }
                     const pg_cfg = this.state.config.pgs.items[pool_id][pg_num];
                     if (pg_cfg)
                     {
@@ -1553,10 +1594,6 @@ class Mon
 
     async apply_pool_pgs(results, up_osds, osd_tree, tree_hash)
     {
-        if (!results.length)
-        {
-            return true;
-        }
         for (const pool_id in (this.state.config.pgs||{}).items||{})
         {
             // We should stop all PGs when deleting a pool or changing its PG count
@@ -1689,7 +1726,6 @@ class Mon
 
     derive_osd_stats(st, prev, prev_diff)
     {
-        const zero_stats = { op: { bps: 0n, iops: 0n, lat: 0n }, subop: { iops: 0n, lat: 0n }, recovery: { bps: 0n, iops: 0n } };
         const diff = { op_stats: {}, subop_stats: {}, recovery_stats: {}, inode_stats: {} };
         if (!st || !st.time || !prev || !prev.time || prev.time >= st.time)
         {
@@ -1729,7 +1765,7 @@ class Mon
         }
         for (const pool_id in st.inode_stats||{})
         {
-            const pool_diff = diff.inode_stats[pool_id] = {};
+            diff.inode_stats[pool_id] = {};
             for (const inode_num in st.inode_stats[pool_id])
             {
                 const inode_diff = diff.inode_stats[pool_id][inode_num] = {};
@@ -2147,7 +2183,7 @@ class Mon
     _die(err, code)
     {
         // In fact we can just try to rejoin
-        console.error(new Error(err || 'Cluster connection failed'));
+        console.error(err instanceof Error ? err : new Error(err || 'Cluster connection failed'));
         process.exit(code || 2);
     }
 
@@ -2171,7 +2207,7 @@ class Mon
 
 function POST(url, body, timeout)
 {
-    return new Promise((ok, no) =>
+    return new Promise(ok =>
     {
         const body_text = Buffer.from(JSON.stringify(body));
         let timer_id = timeout > 0 ? setTimeout(() =>

mon/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor-mon",
-  "version": "1.6.0",
+  "version": "1.6.1",
   "description": "Vitastor SDS monitor service",
   "main": "mon-main.js",
   "scripts": {
@@ -11,5 +11,15 @@
   "dependencies": {
     "sprintf-js": "^1.1.2",
     "ws": "^7.2.5"
+  },
+  "devDependencies": {
+    "eslint": "^8.0.0",
+    "eslint-plugin-node": "^11.1.0"
+  },
+  "engines": {
+    "node": ">=12.0.0"
+  },
+  "scripts": {
+    "lint": "eslint *.js"
   }
 }

mon/simple-offsets.js

@@ -38,7 +38,7 @@ async function run()
         const st = await fs.stat(options.device);
         options.device_block_size = st.blksize;
         if (st.isBlockDevice())
-            device_size = Number(await system("/sbin/blockdev --getsize64 "+options.device))
+            device_size = Number(await system("/sbin/blockdev --getsize64 "+options.device));
         else
             device_size = st.size;
     }
@@ -91,7 +91,7 @@ async function run()
 
 function system(cmd)
 {
-    return new Promise((ok, no) => child_process.exec(cmd, { maxBuffer: 64*1024*1024 }, (err, stdout, stderr) => (err ? no(err.message) : ok(stdout))));
+    return new Promise((ok, no) => child_process.exec(cmd, { maxBuffer: 64*1024*1024 }, (err, stdout/*, stderr*/) => (err ? no(err.message) : ok(stdout))));
 }
 
 run().catch(err => { console.error(err); process.exit(1); });

mon/simple_pgs.js

@@ -198,7 +198,6 @@ function all_combinations(osd_tree, pg_size, ordered, count)
 
 function check_combinations(osd_tree, pgs)
 {
-    const hosts = Object.keys(osd_tree).sort();
     const host_per_osd = {};
     for (const host in osd_tree)
     {
@@ -235,6 +234,7 @@ function compat(params)
 
 module.exports = {
     flatten_tree,
+    all_combinations,
     SimpleCombinator,
     compat,
     NO_OSD,

mon/test-optimize.js

@@ -20,7 +20,7 @@ const osd_tree = {
     },
     500: {
         4: 3.58498,
-//        8: 3.58589,
+        /*8: 3.58589,*/
         9: 3.63869,
     },
     600: {

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VERSION = '1.6.0'
+VERSION = '1.6.1'
 
 LOG = logging.getLogger(__name__)
 

rpm/build-tarball.sh

@@ -24,4 +24,4 @@ rm fio
 mv fio-copy fio
 FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Version[\s:]+(\S+)/ && print $1; /^Release[\s:]+(\S+)/ && print "-$1"; }'`
 perl -i -pe 's/(Requires:\s*fio)([^\n]+)?/$1 = '$FIO'/' $VITASTOR/rpm/vitastor-el$EL.spec
-tar --transform 's#^#vitastor-1.6.0/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-1.6.0$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-1.6.1/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-1.6.1$(rpm --eval '%dist').tar.gz *

rpm/vitastor-el7.Dockerfile

@@ -36,7 +36,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-1.6.0.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-1.6.1.el7.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el7.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        1.6.0
+Version:        1.6.1
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-1.6.0.el7.tar.gz
+Source0:        vitastor-1.6.1.el7.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -104,7 +104,7 @@ rm -rf $RPM_BUILD_ROOT
 %make_install
 . /opt/rh/rh-nodejs12/enable
 cd mon
-npm install
+npm install --production
 cd ..
 mkdir -p %buildroot/usr/lib/vitastor
 cp -r mon %buildroot/usr/lib/vitastor

rpm/vitastor-el8.Dockerfile

@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-1.6.0.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-1.6.1.el8.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el8.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        1.6.0
+Version:        1.6.1
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-1.6.0.el8.tar.gz
+Source0:        vitastor-1.6.1.el8.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -101,7 +101,7 @@ Vitastor fio drivers for benchmarking.
 rm -rf $RPM_BUILD_ROOT
 %make_install
 cd mon
-npm install
+npm install --production
 cd ..
 mkdir -p %buildroot/usr/lib/vitastor
 cp -r mon %buildroot/usr/lib/vitastor

rpm/vitastor-el9.Dockerfile

@@ -18,7 +18,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-1.6.0.el9.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-1.6.1.el9.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el9.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        1.6.0
+Version:        1.6.1
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-1.6.0.el9.tar.gz
+Source0:        vitastor-1.6.1.el9.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -94,7 +94,7 @@ Vitastor fio drivers for benchmarking.
 rm -rf $RPM_BUILD_ROOT
 %cmake_install
 cd mon
-npm install
+npm install --production
 cd ..
 mkdir -p %buildroot/usr/lib/vitastor
 cp -r mon %buildroot/usr/lib/vitastor

src/CMakeLists.txt

@@ -19,7 +19,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 	set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
 endif()
 
-add_definitions(-DVERSION="1.6.0")
+add_definitions(-DVERSION="1.6.1")
 add_definitions(-D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -fno-omit-frame-pointer -I ${CMAKE_SOURCE_DIR}/src)
 add_link_options(-fno-omit-frame-pointer)
 if (${WITH_ASAN})

src/cli_pool_cfg.cpp

@@ -71,7 +71,7 @@ std::string validate_pool_config(json11::Json::object & new_cfg, json11::Json ol
         auto & key = kv_it->first;
         auto & value = kv_it->second;
         if (key == "pg_size" || key == "parity_chunks" || key == "pg_minsize" ||
-            key == "pg_count" || key == "max_osd_combinations" || key == "block_size" ||
+            key == "pg_count" || key == "max_osd_combinations" ||
             key == "bitmap_granularity" || key == "pg_stripe_size")
         {
             if (value.is_number() && value.uint64_value() != value.number_value() ||
@@ -81,6 +81,14 @@ std::string validate_pool_config(json11::Json::object & new_cfg, json11::Json ol
             }
             value = value.uint64_value();
         }
+        else if (key == "block_size")
+        {
+            value = value.is_string() ? parse_size(value.string_value()) : value.uint64_value();
+            if (!value)
+            {
+                return key+" must be an integer with or without size suffix (K/M/G/T)";
+            }
+        }
         else if (key == "name" || key == "scheme" || key == "immediate_commit" ||
             key == "failure_domain" || key == "root_node" || key == "scrub_interval" || key == "used_for_fs" ||
             key == "raw_placement")

src/cli_pool_create.cpp

@@ -14,7 +14,7 @@
 struct pool_creator_t
 {
     cli_tool_t *parent;
-    json11::Json::object cfg;
+    json11::Json cfg;
 
     bool force = false;
     bool wait = false;
@@ -55,8 +55,12 @@ struct pool_creator_t
             goto resume_8;
 
         // Validate pool parameters
-        result.text = validate_pool_config(cfg, json11::Json(), parent->cli->st_cli.global_block_size,
-            parent->cli->st_cli.global_bitmap_granularity, force);
+        {
+            auto new_cfg = cfg.object_items();
+            result.text = validate_pool_config(new_cfg, json11::Json(), parent->cli->st_cli.global_block_size,
+                parent->cli->st_cli.global_bitmap_granularity, force);
+            cfg = new_cfg;
+        }
         if (result.text != "")
         {
             result.err = EINVAL;
@@ -605,7 +609,7 @@ std::function<bool(cli_result_t &)> cli_tool_t::start_pool_create(json11::Json c
 {
     auto pool_creator = new pool_creator_t();
     pool_creator->parent = this;
-    pool_creator->cfg = cfg.object_items();
+    pool_creator->cfg = cfg;
     pool_creator->force = cfg["force"].bool_value();
     pool_creator->wait = cfg["wait"].bool_value();
     return [pool_creator](cli_result_t & result)

src/cli_pool_ls.cpp

@@ -205,7 +205,7 @@ resume_1:
                 { "raw_to_usable", pool_stats[pool_cfg.id]["raw_to_usable"].number_value() },
                 { "space_efficiency", pool_stats[pool_cfg.id]["space_efficiency"].number_value() },
                 { "pg_real_size", pool_stats[pool_cfg.id]["pg_real_size"].uint64_value() },
-                { "osd_count", pg_per_osd.size() },
+                { "osd_count", (uint64_t)pg_per_osd.size() },
             };
         }
         // Include full pool config

src/kv_cli.cpp

@@ -452,7 +452,7 @@ void kv_cli_t::handle_cmd(const std::vector<std::string> & cmd, std::function<vo
         auto name = cmd.size() > 1 ? cmd[1] : "";
         uint64_t pool_id = 0;
         inode_t inode_id = 0;
-        int scanned = sscanf(name.c_str(), "%lu %lu", &pool_id, &inode_id);
+        int scanned = sscanf(name.c_str(), "%ju %ju", &pool_id, &inode_id);
         if (scanned < 2 || !pool_id || !inode_id)
         {
             inode_id = 0;
@@ -483,7 +483,7 @@ void kv_cli_t::handle_cmd(const std::vector<std::string> & cmd, std::function<vo
             else
             {
                 opened = true;
-                fprintf(interactive ? stdout : stderr, "Index opened. Current size: %lu bytes\n", db->get_size());
+                fprintf(interactive ? stdout : stderr, "Index opened. Current size: %ju bytes\n", db->get_size());
             }
             cb(res);
         });

src/kv_db.cpp

@@ -250,14 +250,14 @@ int kv_block_t::parse(uint64_t offset, uint8_t *data, int size)
     {
         // empty block
         if (offset != 0)
-            fprintf(stderr, "K/V: Block %lu is %s\n", offset, blk->magic == 0 ? "empty" : "cleared");
+            fprintf(stderr, "K/V: Block %ju is %s\n", offset, blk->magic == 0 ? "empty" : "cleared");
         return -ENOTBLK;
     }
     if (blk->magic != KV_BLOCK_MAGIC || blk->block_size != size ||
         !blk->type || blk->type > KV_EMPTY || blk->items > KV_BLOCK_MAX_ITEMS)
     {
         // invalid block
-        fprintf(stderr, "K/V: Invalid block %lu magic, size, type or item count\n", offset);
+        fprintf(stderr, "K/V: Invalid block %ju magic, size, type or item count\n", offset);
         return -EILSEQ;
     }
     assert(!this->type);
@@ -266,13 +266,13 @@ int kv_block_t::parse(uint64_t offset, uint8_t *data, int size)
     this->key_ge = read_string(data, size, &pos);
     if (pos < 0)
     {
-        fprintf(stderr, "K/V: Invalid block %lu left bound\n", offset);
+        fprintf(stderr, "K/V: Invalid block %ju left bound\n", offset);
         return -EILSEQ;
     }
     this->key_lt = read_string(data, size, &pos);
     if (pos < 0)
     {
-        fprintf(stderr, "K/V: Invalid block %lu right bound\n", offset);
+        fprintf(stderr, "K/V: Invalid block %ju right bound\n", offset);
         return -EILSEQ;
     }
     if (this->type == KV_INT_SPLIT || this->type == KV_LEAF_SPLIT)
@@ -280,12 +280,12 @@ int kv_block_t::parse(uint64_t offset, uint8_t *data, int size)
         this->right_half = read_string(data, size, &pos);
         if (pos < 0)
         {
-            fprintf(stderr, "K/V: Invalid block %lu split bound\n", offset);
+            fprintf(stderr, "K/V: Invalid block %ju split bound\n", offset);
             return -EILSEQ;
         }
         if (pos+8 > size)
         {
-            fprintf(stderr, "K/V: Invalid block %lu split block ref\n", offset);
+            fprintf(stderr, "K/V: Invalid block %ju split block ref\n", offset);
             return -EILSEQ;
         }
         this->right_half_block = *(uint64_t*)(data+pos);
@@ -296,13 +296,13 @@ int kv_block_t::parse(uint64_t offset, uint8_t *data, int size)
         auto key = read_string(data, size, &pos);
         if (pos < 0)
         {
-            fprintf(stderr, "K/V: Invalid block %lu key %d\n", offset, i);
+            fprintf(stderr, "K/V: Invalid block %ju key %d\n", offset, i);
             return -EILSEQ;
         }
         auto value = read_string(data, size, &pos);
         if (pos < 0)
         {
-            fprintf(stderr, "K/V: Invalid block %lu value %d\n", offset, i);
+            fprintf(stderr, "K/V: Invalid block %ju value %d\n", offset, i);
             return -EILSEQ;
         }
         this->data[key] = value;
@@ -470,7 +470,7 @@ static void dump_str(const std::string & str)
 void kv_block_t::dump(int base_level)
 {
     printf(
-        "{\n    \"block\": %lu,\n    \"level\": %d,\n    \"type\": \"%s\",\n    \"range\": [",
+        "{\n    \"block\": %ju,\n    \"level\": %d,\n    \"type\": \"%s\",\n    \"range\": [",
         offset, base_level+level,
         type < sizeof(block_type_names)/sizeof(block_type_names[0]) ? block_type_names[type] : "unknown"
     );
@@ -482,7 +482,7 @@ void kv_block_t::dump(int base_level)
     {
         printf("    \"right_half\": { ");
         dump_str(right_half);
-        printf(": %lu },\n", right_half_block);
+        printf(": %ju },\n", right_half_block);
     }
     printf("    \"data\": {\n");
     for (auto & kv: data)
@@ -493,7 +493,7 @@ void kv_block_t::dump(int base_level)
         if (type == KV_LEAF || type == KV_LEAF_SPLIT || kv.second.size() != 8)
             dump_str(kv.second);
         else
-            printf("%lu", *(uint64_t*)kv.second.c_str());
+            printf("%ju", *(uint64_t*)kv.second.c_str());
         printf(",\n");
     }
     printf("    }\n}\n");
@@ -1037,7 +1037,7 @@ void kv_op_t::get()
         {
             if (cur_block != 0)
             {
-                fprintf(stderr, "K/V: Hit empty block %lu while searching\n", cur_block);
+                fprintf(stderr, "K/V: Hit empty block %ju while searching\n", cur_block);
                 finish(-EILSEQ);
             }
             else
@@ -1095,7 +1095,7 @@ int kv_op_t::handle_block(int res, int refresh, bool stop_on_split)
         bool fatal = !this->updating_on_path && this->retry > 0;
         if (fatal || db->log_level > 0)
         {
-            fprintf(stderr, "K/V: %sgot unrelated block %lu: key=%s range=[%s, %s) from=[%s, %s)\n",
+            fprintf(stderr, "K/V: %sgot unrelated block %ju: key=%s range=[%s, %s) from=[%s, %s)\n",
                 fatal ? "Error: " : "Warning: read/update collision: ",
                 cur_block, key.c_str(), blk->key_ge.c_str(), blk->key_lt.c_str(), prev_key_ge.c_str(), prev_key_lt.c_str());
             if (fatal)
@@ -1160,7 +1160,7 @@ int kv_op_t::handle_block(int res, int refresh, bool stop_on_split)
         auto child_it = blk->data.upper_bound(key);
         if (child_it == blk->data.begin())
         {
-            fprintf(stderr, "K/V: Internal block %lu misses boundary for %s\n", cur_block, key.c_str());
+            fprintf(stderr, "K/V: Internal block %ju misses boundary for %s\n", cur_block, key.c_str());
             return -EILSEQ;
         }
         auto m = child_it == blk->data.end()
@@ -1169,7 +1169,7 @@ int kv_op_t::handle_block(int res, int refresh, bool stop_on_split)
         child_it--;
         if (child_it->second.size() != sizeof(uint64_t))
         {
-            fprintf(stderr, "K/V: Internal block %lu reference is not 8 byte long\n", cur_block);
+            fprintf(stderr, "K/V: Internal block %ju reference is not 8 byte long\n", cur_block);
             blk->dump(db->base_block_level);
             return -EILSEQ;
         }
@@ -1246,7 +1246,7 @@ static void write_block(kv_db_t *db, kv_block_t *blk, std::function<void(int)> c
         blk->dump(db->base_block_level);
         uint64_t old_size = blk->data_size;
         blk->set_data_size();
-        fprintf(stderr, "K/V: block %lu (ptr=%lx) grew too large: tracked %lu, but real is %u bytes\n",
+        fprintf(stderr, "K/V: block %ju (ptr=%jx) grew too large: tracked %ju, but real is %u bytes\n",
             blk->offset, (uint64_t)blk, old_size, blk->data_size);
         abort();
         return;
@@ -1479,7 +1479,7 @@ void kv_op_t::update_find()
             else if (cur_block == 0)
                 finish(-ENOENT);
             else
-                fprintf(stderr, "K/V: Hit empty block %lu while searching\n", cur_block);
+                fprintf(stderr, "K/V: Hit empty block %ju while searching\n", cur_block);
         }
         else if (res == -ECHILD)
         {
@@ -1506,7 +1506,7 @@ void kv_op_t::create_root()
     // if a referenced non-root block is empty, we just return an error.
     if (cur_block != 0 || db->next_free != 0)
     {
-        fprintf(stderr, "K/V: create_root called with non-empty DB (cur_block=%lu)\n", cur_block);
+        fprintf(stderr, "K/V: create_root called with non-empty DB (cur_block=%ju)\n", cur_block);
         finish(-EILSEQ);
         return;
     }
@@ -1551,7 +1551,7 @@ void kv_op_t::resume_split()
     if (path.size() == 1)
     {
         // It shouldn't be the root block because we don't split it via INT_SPLIT/LEAF_SPLIT
-        fprintf(stderr, "K/V: resume_split at root item (cur_block=%lu)\n", cur_block);
+        fprintf(stderr, "K/V: resume_split at root item (cur_block=%ju)\n", cur_block);
         finish(-EILSEQ);
         return;
     }
@@ -1638,7 +1638,7 @@ void kv_op_t::update_block(int path_pos, bool is_delete, const std::string & key
         {
             blk->dump(0);
             // Should not happen - we should have resumed the split
-            fprintf(stderr, "K/V: attempt to write into block %lu instead of resuming the split (got here from %s..%s)\n",
+            fprintf(stderr, "K/V: attempt to write into block %ju instead of resuming the split (got here from %s..%s)\n",
                 blk->offset, prev_key_ge.c_str(), prev_key_lt.c_str());
             abort();
         }
@@ -1650,7 +1650,7 @@ void kv_op_t::update_block(int path_pos, bool is_delete, const std::string & key
         // No need to split the block => just modify and write it
         if ((blk->type == KV_LEAF_SPLIT || blk->type == KV_INT_SPLIT) && key >= blk->right_half)
         {
-            fprintf(stderr, "K/V: attempt to modify %s in unrelated split block %lu [%s..%s..%s)\n",
+            fprintf(stderr, "K/V: attempt to modify %s in unrelated split block %ju [%s..%s..%s)\n",
                 key.c_str(), blk->offset, blk->key_ge.c_str(), blk->right_half.c_str(), blk->key_lt.c_str());
             blk->dump(db->base_block_level);
             abort();
@@ -1753,11 +1753,11 @@ void kv_op_t::update_block(int path_pos, bool is_delete, const std::string & key
                         clear_block(db, left_blk, 0, [=, left_offset = left_blk->offset](int res)
                         {
                             if (res < 0)
-                                fprintf(stderr, "Failed to clear unreferenced block %lu: %s (code %d)\n", left_offset, strerror(-res), res);
+                                fprintf(stderr, "Failed to clear unreferenced block %ju: %s (code %d)\n", left_offset, strerror(-res), res);
                             clear_block(db, right_blk, 0, [=, right_offset = right_blk->offset](int res)
                             {
                                 if (res < 0)
-                                    fprintf(stderr, "Failed to clear unreferenced block %lu: %s (code %d)\n", right_offset, strerror(-res), res);
+                                    fprintf(stderr, "Failed to clear unreferenced block %ju: %s (code %d)\n", right_offset, strerror(-res), res);
                                 // CAS failure - zero garbage left_blk and right_blk and retry from the beginning
                                 if (write_res == -EINTR)
                                     update();
@@ -1784,7 +1784,7 @@ void kv_op_t::update_block(int path_pos, bool is_delete, const std::string & key
             if (path_pos == 0)
             {
                 // Block number zero should always be the root block
-                fprintf(stderr, "K/V: root block is not 0, but %lu\n", cur_block);
+                fprintf(stderr, "K/V: root block is not 0, but %ju\n", cur_block);
                 cb(-EILSEQ);
                 return;
             }
@@ -1809,7 +1809,7 @@ void kv_op_t::update_block(int path_pos, bool is_delete, const std::string & key
                     clear_block(db, right_blk, 0, [=, right_offset = right_blk->offset](int res)
                     {
                         if (res < 0)
-                            fprintf(stderr, "Failed to clear unreferenced block %lu: %s (code %d)\n", right_offset, strerror(-res), res);
+                            fprintf(stderr, "Failed to clear unreferenced block %ju: %s (code %d)\n", right_offset, strerror(-res), res);
                         // CAS failure - zero garbage right_blk and retry from the beginning
                         if (write_res == -EINTR)
                             update();
@@ -1860,7 +1860,7 @@ void kv_op_t::next_handle_block(int res, int refresh)
             finish(-ENOENT);
         else
         {
-            fprintf(stderr, "K/V: Hit empty block %lu while searching\n", cur_block);
+            fprintf(stderr, "K/V: Hit empty block %ju while searching\n", cur_block);
             finish(-EILSEQ);
         }
     }

src/kv_stress.cpp

@@ -632,7 +632,7 @@ void kv_test_t::print_stats(kv_test_stat_t & prev_stat, timespec & prev_stat_tim
             char buf[128] = { 0 };
             for (int i = 0; i < sizeof(lats)/sizeof(lats[0]); i++)
             {
-                snprintf(buf, sizeof(buf)-1, "%.1f %s/s (%lu us)", (lats[i]->count-prev[i]->count)*1000000.0/usec,
+                snprintf(buf, sizeof(buf)-1, "%.1f %s/s (%ju us)", (lats[i]->count-prev[i]->count)*1000000.0/usec,
                     lats[i]->name, (lats[i]->usec-prev[i]->usec)/(lats[i]->count-prev[i]->count > 0 ? lats[i]->count-prev[i]->count : 1));
                 int k;
                 for (k = strlen(buf); k < strlen(lats[i]->name)+21; k++)
@@ -652,7 +652,7 @@ void kv_test_t::print_stats(kv_test_stat_t & prev_stat, timespec & prev_stat_tim
                 if (lats[i]->count > prev[i]->count)
                 {
                     printf(
-                        ",\"%s\":{\"avg\":{\"iops\":%.1f,\"usec\":%lu},\"total\":{\"count\":%lu,\"usec\":%lu}}",
+                        ",\"%s\":{\"avg\":{\"iops\":%.1f,\"usec\":%ju},\"total\":{\"count\":%ju,\"usec\":%ju}}",
                         lats[i]->name, (lats[i]->count-prev[i]->count)*1000000.0/usec,
                         (lats[i]->usec-prev[i]->usec)/(lats[i]->count-prev[i]->count),
                         lats[i]->count, lats[i]->usec

src/nfs_block.cpp

@@ -497,7 +497,7 @@ static void extend_inode(nfs_client_t *self, uint64_t inode, uint64_t new_size)
             auto & ext = self->parent->blockfs->extends[inode];
             if (r.err)
             {
-                fprintf(stderr, "Error extending inode %lu to %lu bytes: %s\n", inode, new_size, r.text.c_str());
+                fprintf(stderr, "Error extending inode %ju to %ju bytes: %s\n", inode, new_size, r.text.c_str());
             }
             if (r.err == EAGAIN || ext.next_extend > ext.cur_extend)
             {

src/nfs_kv_readdir.cpp

@@ -258,7 +258,7 @@ resume_3:
         auto name = kv_direntry_filename(st->cur_key);
         if (st->self->parent->trace)
         {
-            fprintf(stderr, "[%d] READDIR %ju %lu %s\n",
+            fprintf(stderr, "[%d] READDIR %ju %ju %s\n",
                 st->self->nfs_fd, st->dir_ino, st->offset, name.c_str());
         }
         auto fh = kv_fh(ino);

src/nfs_kv_write.cpp

@@ -608,7 +608,7 @@ static void nfs_kv_extend_inode(nfs_kv_write_state *st, int state, int base_stat
         auto ientry = json11::Json::parse(old_value, err).object_items();
         if (err != "")
         {
-            fprintf(stderr, "Invalid JSON in inode %lu = %s: %s\n", st->ino, old_value.c_str(), err.c_str());
+            fprintf(stderr, "Invalid JSON in inode %ju = %s: %s\n", st->ino, old_value.c_str(), err.c_str());
             st->res2 = -EINVAL;
             return false;
         }

src/osd_cluster.cpp

@@ -534,7 +534,7 @@ void osd_t::renew_lease(bool reload)
         { "ID", etcd_lease_id }
     }, st_cli.etcd_quick_timeout, 0, 0, [this, reload](std::string err, json11::Json data)
     {
-        if (err == "" && data["result"]["TTL"].string_value() == "")
+        if (err == "" && data["result"]["TTL"].uint64_value() == 0)
         {
             // Die
             fprintf(stderr, "Error refreshing etcd lease\n");

src/vitastor.pc.in

@@ -6,7 +6,7 @@ includedir=${prefix}/@CMAKE_INSTALL_INCLUDEDIR@
 
 Name: Vitastor
 Description: Vitastor client library
-Version: 1.6.0
+Version: 1.6.1
 Libs: -L${libdir} -lvitastor_client
 Cflags: -I${includedir}