Release 0.5.8

- Add heartbeats (fixes failover in case of network issues or offline nodes)
- Fix a bug where a PG could incorrectly become listed as 'incomplete' if historical osd_sets
  included a set with the the PG's primary OSD as the only alive one
- Use osd_out_time = 10 minutes by default instead of 30 minutes
- Make monitors stick to a single selected etcd URL on start and not try to select random ones
  on every request - this was leading to etcd interaction errors when some etcds were unavailable

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (0.5.7-1) unstable; urgency=medium
+vitastor (0.5.8-1) unstable; urgency=medium
 
   * Bugfixes
 

debian/vitastor.Dockerfile

@@ -40,10 +40,10 @@ RUN set -e -x; \
     mkdir -p /root/packages/vitastor-$REL; \
     rm -rf /root/packages/vitastor-$REL/*; \
     cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-0.5.7; \
-    ln -s /root/packages/qemu-$REL/qemu-*/ vitastor-0.5.7/qemu; \
-    ln -s /root/fio-build/fio-*/ vitastor-0.5.7/fio; \
-    cd vitastor-0.5.7; \
+    cp -r /root/vitastor vitastor-0.5.8; \
+    ln -s /root/packages/qemu-$REL/qemu-*/ vitastor-0.5.8/qemu; \
+    ln -s /root/fio-build/fio-*/ vitastor-0.5.8/fio; \
+    cd vitastor-0.5.8; \
     FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     QEMU=$(head -n1 qemu/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     sh copy-qemu-includes.sh; \
@@ -59,8 +59,8 @@ RUN set -e -x; \
     echo "dep:fio=$FIO" > debian/substvars; \
     echo "dep:qemu=$QEMU" >> debian/substvars; \
     cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.5.7.orig.tar.xz vitastor-0.5.7; \
-    cd vitastor-0.5.7; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.5.8.orig.tar.xz vitastor-0.5.8; \
+    cd vitastor-0.5.8; \
     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; \

mon/mon.js

@@ -35,7 +35,7 @@ const etcd_tree = {
             etcd_mon_retries: 5, // min: 0
             mon_change_timeout: 1000, // ms. min: 100
             mon_stats_timeout: 1000, // ms. min: 100
-            osd_out_time: 1800, // seconds. min: 0
+            osd_out_time: 600, // seconds. min: 0
             placement_levels: { datacenter: 1, rack: 2, host: 3, osd: 4, ... },
             // client and osd
             use_sync_send_recv: false,
@@ -295,7 +295,7 @@ class Mon
         this.config.osd_out_time = Number(this.config.osd_out_time) || 0;
         if (!this.config.osd_out_time)
         {
-            this.config.osd_out_time = 30*60; // 30 minutes by default
+            this.config.osd_out_time = 600; // 10 minutes by default
         }
     }
 
@@ -317,8 +317,14 @@ class Mon
                     ok(false);
                 }, this.config.etcd_mon_timeout);
                 this.ws = new WebSocket(base+'/watch');
+                const fail = () =>
+                {
+                    ok(false);
+                };
+                this.ws.on('error', fail);
                 this.ws.on('open', () =>
                 {
+                    this.ws.removeListener('error', fail);
                     if (timer_id)
                         clearTimeout(timer_id);
                     ok(true);
@@ -445,7 +451,7 @@ class Mon
     async get_lease()
     {
         const max_ttl = this.config.etcd_mon_ttl + this.config.etcd_mon_timeout/1000*this.config.etcd_mon_retries;
-        const res = await this.etcd_call('/lease/grant', { TTL: max_ttl }, this.config.etcd_mon_timeout, this.config.etcd_mon_retries);
+        const res = await this.etcd_call('/lease/grant', { TTL: max_ttl }, this.config.etcd_mon_timeout, -1);
         this.etcd_lease_id = res.ID;
         setInterval(async () =>
         {
@@ -1241,6 +1247,11 @@ class Mon
                     console.error('etcd returned error: '+res.json.error);
                     break;
                 }
+                if (this.etcd_urls.length > 1)
+                {
+                    // Stick to the same etcd for the rest of calls
+                    this.etcd_urls = [ base ];
+                }
                 return res.json;
             }
             retry++;

rpm/build-tarball.sh

@@ -48,4 +48,4 @@ FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Ve
 QEMU=`rpm -qi qemu qemu-kvm | 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
 perl -i -pe 's/(Requires:\s*qemu(?:-kvm)?)([^\n]+)?/$1 = '$QEMU'/' $VITASTOR/rpm/vitastor-el$EL.spec
-tar --transform 's#^#vitastor-0.5.7/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.5.7$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-0.5.8/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.5.8$(rpm --eval '%dist').tar.gz *

rpm/vitastor-el7.Dockerfile

@@ -37,7 +37,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-0.5.7.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.5.8.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:        0.5.7
+Version:        0.5.8
 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-0.5.7.el7.tar.gz
+Source0:        vitastor-0.5.8.el7.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

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-0.5.7.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.5.8.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:        0.5.7
+Version:        0.5.8
 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-0.5.7.el8.tar.gz
+Source0:        vitastor-0.5.8.el8.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel

src/cluster_client.cpp

@@ -8,13 +8,11 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
 {
     this->ringloop = ringloop;
     this->tfd = tfd;
-
-    log_level = config["log_level"].int64_value();
+    this->config = config;
 
     msgr.osd_num = 0;
     msgr.tfd = tfd;
     msgr.ringloop = ringloop;
-    msgr.log_level = log_level;
     msgr.repeer_pgs = [this](osd_num_t peer_osd)
     {
         if (msgr.osd_peer_fds.find(peer_osd) != msgr.osd_peer_fds.end())
@@ -67,8 +65,7 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
         msgr.stop_client(op->peer_fd);
         delete op;
     };
-    msgr.use_sync_send_recv = config["use_sync_send_recv"].bool_value() ||
-        config["use_sync_send_recv"].uint64_value();
+    msgr.init();
 
     st_cli.tfd = tfd;
     st_cli.on_load_config_hook = [this](json11::Json::object & cfg) { on_load_config_hook(cfg); };
@@ -185,16 +182,8 @@ void cluster_client_t::on_load_config_hook(json11::Json::object & config)
     {
         up_wait_retry_interval = 50;
     }
-    msgr.peer_connect_interval = config["peer_connect_interval"].uint64_value();
-    if (!msgr.peer_connect_interval)
-    {
-        msgr.peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
-    }
-    msgr.peer_connect_timeout = config["peer_connect_timeout"].uint64_value();
-    if (!msgr.peer_connect_timeout)
-    {
-        msgr.peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
-    }
+    msgr.parse_config(config);
+    msgr.parse_config(this->config);
     st_cli.load_pgs();
 }
 

src/cluster_client.h

@@ -82,6 +82,7 @@ class cluster_client_t
 public:
     etcd_state_client_t st_cli;
     osd_messenger_t msgr;
+    json11::Json config;
 
     cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json & config);
     ~cluster_client_t();

src/messenger.cpp

@@ -26,14 +26,106 @@ osd_op_t::~osd_op_t()
     }
 }
 
+void osd_messenger_t::init()
+{
+    keepalive_timer_id = tfd->set_timer(1000, true, [this](int)
+    {
+        for (auto cl_it = clients.begin(); cl_it != clients.end();)
+        {
+            auto cl = (cl_it++)->second;
+            if (!cl->osd_num)
+            {
+                // Do not run keepalive on regular clients
+                continue;
+            }
+            if (cl->ping_time_remaining > 0)
+            {
+                cl->ping_time_remaining--;
+                if (!cl->ping_time_remaining)
+                {
+                    // Ping timed out, stop the client
+                    stop_client(cl->peer_fd, true);
+                }
+            }
+            else if (cl->idle_time_remaining > 0)
+            {
+                cl->idle_time_remaining--;
+                if (!cl->idle_time_remaining)
+                {
+                    // Connection is idle for <osd_idle_time>, send ping
+                    osd_op_t *op = new osd_op_t();
+                    op->op_type = OSD_OP_OUT;
+                    op->peer_fd = cl->peer_fd;
+                    op->req = (osd_any_op_t){
+                        .hdr = {
+                            .magic = SECONDARY_OSD_OP_MAGIC,
+                            .id = this->next_subop_id++,
+                            .opcode = OSD_OP_PING,
+                        },
+                    };
+                    op->callback = [this, cl](osd_op_t *op)
+                    {
+                        int fail_fd = (op->reply.hdr.retval != 0 ? op->peer_fd : -1);
+                        cl->ping_time_remaining = 0;
+                        delete op;
+                        if (fail_fd >= 0)
+                        {
+                            stop_client(fail_fd, true);
+                        }
+                    };
+                    outbox_push(op);
+                    cl->ping_time_remaining = osd_ping_timeout;
+                    cl->idle_time_remaining = osd_idle_timeout;
+                }
+            }
+            else
+            {
+                cl->idle_time_remaining = osd_idle_timeout;
+            }
+        }
+    });
+}
+
 osd_messenger_t::~osd_messenger_t()
 {
+    if (keepalive_timer_id >= 0)
+    {
+        tfd->clear_timer(keepalive_timer_id);
+        keepalive_timer_id = -1;
+    }
     while (clients.size() > 0)
     {
         stop_client(clients.begin()->first, true);
     }
 }
 
+void osd_messenger_t::parse_config(const json11::Json & config)
+{
+    this->use_sync_send_recv = config["use_sync_send_recv"].bool_value() ||
+        config["use_sync_send_recv"].uint64_value();
+    this->peer_connect_interval = config["peer_connect_interval"].uint64_value();
+    if (!this->peer_connect_interval)
+    {
+        this->peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
+    }
+    this->peer_connect_timeout = config["peer_connect_timeout"].uint64_value();
+    if (!this->peer_connect_timeout)
+    {
+        this->peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
+    }
+    this->osd_idle_timeout = config["osd_idle_timeout"].uint64_value();
+    if (!this->osd_idle_timeout)
+    {
+        this->osd_idle_timeout = DEFAULT_OSD_PING_TIMEOUT;
+    }
+    this->osd_ping_timeout = config["osd_ping_timeout"].uint64_value();
+    if (!this->osd_ping_timeout)
+    {
+        this->osd_ping_timeout = DEFAULT_OSD_PING_TIMEOUT;
+    }
+    this->log_level = config["log_level"].uint64_value();
+}
+
 void osd_messenger_t::connect_peer(uint64_t peer_osd, json11::Json peer_state)
 {
     if (wanted_peers.find(peer_osd) == wanted_peers.end())

src/messenger.h

@@ -34,6 +34,7 @@
 
 #define DEFAULT_PEER_CONNECT_INTERVAL 5
 #define DEFAULT_PEER_CONNECT_TIMEOUT 5
+#define DEFAULT_OSD_PING_TIMEOUT 5
 
 // Kind of a vector with small-list-optimisation
 struct osd_op_buf_list_t
@@ -198,6 +199,8 @@ struct osd_client_t
     int peer_fd;
     int peer_state;
     int connect_timeout_id = -1;
+    int ping_time_remaining = 0;
+    int idle_time_remaining = 0;
     osd_num_t osd_num = 0;
 
     void *in_buf = NULL;
@@ -251,6 +254,7 @@ struct osd_messenger_t
 {
     timerfd_manager_t *tfd;
     ring_loop_t *ringloop;
+    int keepalive_timer_id = -1;
 
     // osd_num_t is only for logging and asserts
     osd_num_t osd_num;
@@ -258,6 +262,8 @@ struct osd_messenger_t
     int receive_buffer_size = 64*1024;
     int peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
     int peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
+    int osd_idle_timeout = DEFAULT_OSD_PING_TIMEOUT;
+    int osd_ping_timeout = DEFAULT_OSD_PING_TIMEOUT;
     int log_level = 0;
     bool use_sync_send_recv = false;
 
@@ -274,6 +280,8 @@ struct osd_messenger_t
     osd_op_stats_t stats;
 
 public:
+    void init();
+    void parse_config(const json11::Json & config);
     void connect_peer(uint64_t osd_num, json11::Json peer_state);
     void stop_client(int peer_fd, bool force = false);
     void outbox_push(osd_op_t *cur_op);

src/osd.cpp

@@ -37,6 +37,7 @@ osd_t::osd_t(blockstore_config_t & config, blockstore_t *bs, ring_loop_t *ringlo
     c_cli.ringloop = this->ringloop;
     c_cli.exec_op = [this](osd_op_t *op) { exec_op(op); };
     c_cli.repeer_pgs = [this](osd_num_t peer_osd) { repeer_pgs(peer_osd); };
+    c_cli.init();
 
     init_cluster();
 
@@ -100,14 +101,7 @@ void osd_t::parse_config(blockstore_config_t & config)
     slow_log_interval = strtoull(config["slow_log_interval"].c_str(), NULL, 10);
     if (!slow_log_interval)
         slow_log_interval = 10;
-    c_cli.peer_connect_interval = strtoull(config["peer_connect_interval"].c_str(), NULL, 10);
-    if (!c_cli.peer_connect_interval)
-        c_cli.peer_connect_interval = DEFAULT_PEER_CONNECT_INTERVAL;
-    c_cli.peer_connect_timeout = strtoull(config["peer_connect_timeout"].c_str(), NULL, 10);
-    if (!c_cli.peer_connect_timeout)
-        c_cli.peer_connect_timeout = DEFAULT_PEER_CONNECT_TIMEOUT;
-    log_level = strtoull(config["log_level"].c_str(), NULL, 10);
-    c_cli.log_level = log_level;
+    c_cli.parse_config(json_config);
 }
 
 void osd_t::bind_socket()
@@ -211,6 +205,12 @@ void osd_t::exec_op(osd_op_t *cur_op)
         finish_op(cur_op, -EINVAL);
         return;
     }
+    if (cur_op->req.hdr.opcode == OSD_OP_PING)
+    {
+        // Pong
+        finish_op(cur_op, 0);
+        return;
+    }
     if (readonly &&
         cur_op->req.hdr.opcode != OSD_OP_SEC_READ &&
         cur_op->req.hdr.opcode != OSD_OP_SEC_LIST &&

src/osd_ops.cpp

@@ -19,4 +19,5 @@ const char* osd_op_names[] = {
     "primary_write",
     "primary_sync",
     "primary_delete",
+    "ping",
 };

src/osd_ops.h

@@ -27,7 +27,8 @@
 #define OSD_OP_WRITE                12
 #define OSD_OP_SYNC                 13
 #define OSD_OP_DELETE               14
-#define OSD_OP_MAX                  14
+#define OSD_OP_PING                 15
+#define OSD_OP_MAX                  15
 // Alignment & limit for read/write operations
 #ifndef MEM_ALIGNMENT
 #define MEM_ALIGNMENT               512

src/osd_peering.cpp

@@ -183,7 +183,8 @@ void osd_t::start_pg_peering(pg_t & pg)
             bool found = false;
             for (auto history_osd: history_set)
             {
-                if (history_osd != 0 && c_cli.osd_peer_fds.find(history_osd) != c_cli.osd_peer_fds.end())
+                if (history_osd != 0 && (history_osd == this->osd_num ||
+                    c_cli.osd_peer_fds.find(history_osd) != c_cli.osd_peer_fds.end()))
                 {
                     found = true;
                     break;