Add a minimum interval for etcd_state_client to reload state

(To prevent excessive load on etcd during outages)

docs/config/network.en.md

@@ -30,6 +30,7 @@ between clients, OSDs and etcd.
 - [etcd_slow_timeout](#etcd_slow_timeout)
 - [etcd_keepalive_timeout](#etcd_keepalive_timeout)
 - [etcd_ws_keepalive_interval](#etcd_ws_keepalive_interval)
+- [etcd_min_reload_interval](#etcd_min_reload_interval)
 
 ## tcp_header_buffer_size
 
@@ -261,3 +262,13 @@ etcd_report_interval to guarantee that keepalive actually works.
 
 etcd websocket ping interval required to keep the connection alive and
 detect disconnections quickly.
+
+## etcd_min_reload_interval
+
+- Type: milliseconds
+- Default: 1000
+- Can be changed online: yes
+
+Minimum interval for full etcd state reload. Introduced to prevent
+excessive load on etcd during outages when etcd can't keep up with event
+streams and cancels them.

docs/config/network.ru.md

@@ -30,6 +30,7 @@
 - [etcd_slow_timeout](#etcd_slow_timeout)
 - [etcd_keepalive_timeout](#etcd_keepalive_timeout)
 - [etcd_ws_keepalive_interval](#etcd_ws_keepalive_interval)
+- [etcd_min_reload_interval](#etcd_min_reload_interval)
 
 ## tcp_header_buffer_size
 
@@ -271,3 +272,13 @@ etcd_report_interval, чтобы keepalive гарантированно рабо
 - Можно менять на лету: да
 
 Интервал проверки живости вебсокет-подключений к etcd.
+
+## etcd_min_reload_interval
+
+- Тип: миллисекунды
+- Значение по умолчанию: 1000
+- Можно менять на лету: да
+
+Минимальный интервал полной перезагрузки состояния из etcd. Добавлено для
+предотвращения избыточной нагрузки на etcd во время отказов, когда etcd не
+успевает рассылать потоки событий и отменяет их.

docs/config/src/network.yml

@@ -306,3 +306,15 @@
     detect disconnections quickly.
   info_ru: |
     Интервал проверки живости вебсокет-подключений к etcd.
+- name: etcd_min_reload_interval
+  type: ms
+  default: 1000
+  online: true
+  info: |
+    Minimum interval for full etcd state reload. Introduced to prevent
+    excessive load on etcd during outages when etcd can't keep up with event
+    streams and cancels them.
+  info_ru: |
+    Минимальный интервал полной перезагрузки состояния из etcd. Добавлено для
+    предотвращения избыточной нагрузки на etcd во время отказов, когда etcd не
+    успевает рассылать потоки событий и отменяет их.

docs/usage/cli.ru.md

@@ -22,6 +22,8 @@ vitastor-cli - интерфейс командной строки для адм
 - [flatten](#flatten)
 - [rm-data](#rm-data)
 - [merge-data](#merge-data)
+- [describe](#describe)
+- [fix](#fix)
 - [alloc-osd](#alloc-osd)
 - [rm-osd](#rm-osd)
 - [osd-tree](#osd-tree)

src/client/etcd_state_client.cpp

@@ -31,6 +31,11 @@ etcd_state_client_t::~etcd_state_client_t()
         keepalive_client = NULL;
     }
 #endif
+    if (load_pgs_timer_id >= 0)
+    {
+        tfd->clear_timer(load_pgs_timer_id);
+        load_pgs_timer_id = -1;
+    }
 }
 
 #ifndef __MOCK__
@@ -143,6 +148,7 @@ void etcd_state_client_t::etcd_call(std::string api, json11::Json payload, int t
                 }
                 if (interval > 0)
                 {
+                    // FIXME: Prevent destruction of etcd_state_client if timers or requests are active
                     tfd->set_timer(interval, false, [this, api, payload, timeout, retries, interval, callback](int)
                     {
                         etcd_call(api, payload, timeout, retries, interval, callback);
@@ -271,6 +277,11 @@ void etcd_state_client_t::parse_config(const json11::Json & config)
     {
         this->etcd_quick_timeout = 1000;
     }
+    this->etcd_min_reload_interval = config["etcd_min_reload_interval"].uint64_value();
+    if (this->etcd_min_reload_interval <= 0)
+    {
+        this->etcd_min_reload_interval = 50;
+    }
     if (this->etcd_ws_keepalive_interval != old_etcd_ws_keepalive_interval && ws_keepalive_timer >= 0)
     {
 #ifndef __MOCK__
@@ -603,6 +614,23 @@ void etcd_state_client_t::load_global_config()
 
 void etcd_state_client_t::load_pgs()
 {
+    timespec tv;
+    clock_gettime(CLOCK_REALTIME, &tv);
+    uint64_t ms_passed = (tv.tv_sec-etcd_last_reload.tv_sec)*1000 + (tv.tv_nsec-etcd_last_reload.tv_nsec)/1000000;
+    if (ms_passed < etcd_min_reload_interval)
+    {
+        if (load_pgs_timer_id < 0)
+        {
+            load_pgs_timer_id = tfd->set_timer(etcd_min_reload_interval+50-ms_passed, false, [this](int) { load_pgs(); });
+        }
+        return;
+    }
+    etcd_last_reload = tv;
+    if (load_pgs_timer_id >= 0)
+    {
+        tfd->clear_timer(load_pgs_timer_id);
+        load_pgs_timer_id = -1;
+    }
     json11::Json::array txn = {
         json11::Json::object {
             { "request_range", json11::Json::object {

src/client/etcd_state_client.h

@@ -108,6 +108,7 @@ public:
     int max_etcd_attempts = 5;
     int etcd_quick_timeout = 1000;
     int etcd_slow_timeout = 5000;
+    int etcd_min_reload_interval = 1000;
     bool infinite_start = true;
     uint64_t global_block_size = DEFAULT_BLOCK_SIZE;
     uint32_t global_bitmap_granularity = DEFAULT_BITMAP_GRANULARITY;
@@ -122,6 +123,8 @@ public:
     uint64_t etcd_watch_revision_config = 0;
     uint64_t etcd_watch_revision_osd = 0;
     uint64_t etcd_watch_revision_pg = 0;
+    timespec etcd_last_reload = {};
+    int load_pgs_timer_id = -1;
     std::map<pool_id_t, pool_config_t> pool_config;
     std::map<osd_num_t, json11::Json> peer_states;
     std::set<osd_num_t> seen_peers;