View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000193 | Pgpool-II | Bug | public | 2016-05-05 20:55 | 2016-05-21 05:04 |
| Reporter | eldad | Assigned To | |||
| Priority | high | Severity | major | Reproducibility | sometimes |
| Status | closed | Resolution | open | ||
| Platform | Linux | OS | RHEL | OS Version | 6.3 |
| Summary | 0000193: pgpool trigger failover after killing user session in the master postgres server | ||||
| Description | We had a stuck session from application on the master that executed a select. I killed the session using pg_terminate_backend. few seconds after it the pgpool received degenerate backend request for the master and activated failover. It happened 2 time in the last month. | ||||
| Steps To Reproduce | kill active session on the master postgres. | ||||
| Additional Information | 2016-05-05 03:58:20: pid 29211: LOG: pool_send_and_wait: Error or notice message from backend: : DB node id: 2 backend pid: 23748 statement: "SELECT ts, metric_val,script_id,transaction_name, CASE WHEN false OR cnt<=25 THEN lower ELSE NULL END AS lower, CASE WHEN false OR cnt<=25 .... terminating connection due to administrator command" 2016-05-05 03:58:20: pid 29211: ERROR: unable to forward message to frontend 2016-05-05 03:58:20: pid 29211: DETAIL: FATAL error occured on backend 2016-05-05 03:58:21: pid 29211: LOG: received degenerate backend request for node_id: 2 from pid [29211] 2016-05-05 03:58:21: pid 30772: LOG: watchdog notifying to start interlocking 2016-05-05 03:58:21: pid 30772: LOG: watchdog became a new lock holder 2016-05-05 03:58:21: pid 30775: LOG: sending watchdog response 2016-05-05 03:58:21: pid 30775: DETAIL: WD_STAND_FOR_LOCK_HOLDER received but lock holder already exists 2016-05-05 03:58:21: pid 30772: LOG: starting degeneration. shutdown host host12(5432) 2016-05-05 03:58:21: pid 30772: LOG: Restart all children | ||||
| Tags | No tags attached. | ||||
