Fix for no primary on standby pgpool when primary is quarantined on master

Master watchdog Pgpool sends primary_node_id = -1 in the backend status sync
message if the primary node is quarantined on it. So standby watchdog Pgpool
must not update its primary_node_id if the primary backend node id in sync
message is invalid_node_id (-1) while the same sync message reports the
backend status of the current primary node as "NOT DOWN".

The issue was reported by  "Tatsuo Ishii <ishii@sraoss.co.jp>" and fixed by me

diff --git a/src/main/pgpool_main.c b/src/main/pgpool_main.c
index 371aa69cf6818151bdea1083241e35754db488e0..1a03927bced31f65178304ea2117af95850acecd 100644 (file)
--- a/src/main/pgpool_main.c
+++ b/src/main/pgpool_main.c
@@ -4151,17 +4151,6 @@ sync_backend_from_watchdog(void)
        ereport(DEBUG1,
                        (errmsg("primary node on master watchdog node \"%s\" is %d", backendStatus->nodeName, backendStatus->primary_node_id)));
 
-       if (Req_info->primary_node_id != backendStatus->primary_node_id)
-       {
-               /* Do not produce this log message if we are starting up the Pgpool-II */
-               if (processState != INITIALIZING)
-                       ereport(LOG,
-                                       (errmsg("primary node:%d on master watchdog node \"%s\" is different from local primary node:%d",
-                                                       backendStatus->primary_node_id, backendStatus->nodeName, Req_info->primary_node_id)));
-
-               Req_info->primary_node_id = backendStatus->primary_node_id;
-               primary_changed = true;
-       }
 
        /*
         * update the local backend status Also remove quarantine flags
@@ -4204,6 +4193,34 @@ sync_backend_from_watchdog(void)
                        }
                }
        }
+
+       if (Req_info->primary_node_id != backendStatus->primary_node_id)
+       {
+               /* Do not produce this log message if we are starting up the Pgpool-II */
+               if (processState != INITIALIZING)
+                       ereport(LOG,
+                                       (errmsg("primary node:%d on master watchdog node \"%s\" is different from local primary node:%d",
+                                                       backendStatus->primary_node_id, backendStatus->nodeName, Req_info->primary_node_id)));
+               /*
+                * master node returns primary_node_id = -1 when the node primary
+                * node is in  quarantine state on the master.
+                * So we will not update our primary node id when the status of current primary node
+                * is not CON_DOWN while primary_node_id sent by master watchdong node is -1
+                */
+               if (backendStatus->primary_node_id == -1 && BACKEND_INFO(Req_info->primary_node_id).backend_status != CON_DOWN)
+               {
+                       ereport(LOG,
+                (errmsg("primary node:%d on master watchdog node \"%s\" seems to be quarantined",
+                                       Req_info->primary_node_id, backendStatus->nodeName),
+                errdetail("keeping the current primary")));
+               }
+               else
+               {
+                       Req_info->primary_node_id = backendStatus->primary_node_id;
+                       primary_changed = true;
+               }
+       }
+
        pfree(backendStatus);
 
        if (reload_maste_node_id)

diff --git a/src/watchdog/wd_json_data.c b/src/watchdog/wd_json_data.c
index 7f3cb66257f91fc91e7b3ff47775ea7ddc94a11c..4fa8dd77a379aa68593a4bdb35849307155e3cdd 100644 (file)
--- a/src/watchdog/wd_json_data.c
+++ b/src/watchdog/wd_json_data.c
@@ -324,7 +324,7 @@ get_backend_node_status_json(WatchdogNode * wdNode)
                if (backend_status == CON_DOWN && pool_config->backend_desc->backend_info[i].quarantine)
                {
                        /*
-                        * since quarantine nodes are not cluster wide so send CON_WATI
+                        * since quarantine nodes are not cluster wide so send CON_WAIT
                         * status for quarantine nodes
                         */
                        backend_status = CON_CONNECT_WAIT;