View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000454 | Pgpool-II | Bug | public | 2019-01-05 00:05 | 2019-02-18 14:31 |
| Reporter | jprieto | Assigned To | pengbo | ||
| Priority | urgent | Severity | feature | Reproducibility | always |
| Status | closed | Resolution | open | ||
| Product Version | 4.0.2 | ||||
| Summary | 0000454: When multiple failover happens the pgpool pool node status is going into standby -stanby state | ||||
| Description | I have 2 cluster nodes installed with pgpool-II(pgpool-II-4.0.2) and postgresql(postgresql PostgreSQL 9.4.20). The two nodes are configured as streaming replication mode. Lets name the two clusetr node as master-server(primary) and slave-server(secondary ). Use case: 1. Failover from primary(master-server) to secondary(slave-server) 2. Failover from secondary(slave-server) to primary(master-server) Thsi scenario leads the pgpool node status to standby-standby state | ||||
| Steps To Reproduce | 1. Started postgers DB on primary node (master-server) 2. Started the pgpool on primary node (slave-server) 3. Once the delegate IP is up on the primary node (master-server), started the pgpool on the secondary node (slave-serve). 4. pgpool node status is follows postgres=# show pool_nodes; node_id | hostname | port | status | lb_weight | role | select_cnt | load_balance_node | replication_delay | last_status_change ---------+---------------+------+--------+-----------+---------+------------+-------------------+-------------------+--------------------- 0 | master-server | 5432 | up | 0.500000 | primary | 0 | false | 0 | 2019-01-04 10:03:24 1 | slave-server | 5432 | up | 0.500000 | standby | 0 | true | 0 | 2019-01-04 10:57:08 (2 filas) 5. I want to do the failover from master-server to slave-server. SO i brought down the postgres DB on the primary(vm1).Output follows postgres=# show pool_nodes; node_id | hostname | port | status | lb_weight | role | select_cnt | load_balance_node | replication_delay | last_status_change ---------+---------------+------+--------+-----------+---------+------------+-------------------+-------------------+--------------------- 0 | master-server | 5432 | down | 0.500000 | standby | 1 | false | 0 | 2019-01-04 11:13:50 1 | slave-server | 5432 | up | 0.500000 | primary | 1 | true | 0 | 2019-01-04 11:13:50 (2 filas) The secondary(slave-server) node became the primary after the failover. 6. Then started the pcp recovery for primary node(master-server). Output follows [root@slave-server data]# pcp_recovery_node -h 172.20.10.11 -p 9898 -U postgres -n 0 -d postgres=# show pool_nodes; node_id | hostname | port | status | lb_weight | role | select_cnt | load_balance_node | replication_delay | last_status_change ---------+---------------+------+--------+-----------+---------+------------+-------------------+-------------------+--------------------- 0 | master-server | 5432 | up | 0.500000 | standby | 1 | true | 0 | 2019-01-04 11:23:35 1 | slave-server | 5432 | up | 0.500000 | primary | 1 | false | 0 | 2019-01-04 11:13:50 (2 filas) Prmary(master-server) node db is up now. 7. Again i want to do the failover from slave-server to master-server. SO i brought down the DB from slave-server. Output as follows postgres=# show pool_nodes; node_id | hostname | port | status | lb_weight | role | select_cnt | load_balance_node | replication_delay | last_status_change ---------+---------------+------+--------+-----------+---------+------------+-------------------+-------------------+--------------------- 0 | master-server | 5432 | up | 0.500000 | standby | 2 | true | 0 | 2019-01-04 11:23:35 1 | slave-server | 5432 | down | 0.500000 | standby | 1 | false | 0 | 2019-01-04 11:30:45 (2 filas) DOing this leads the pgpool node into standby-standby state. | ||||
| Tags | No tags attached. | ||||
|
|
|
|
|
It seems that Pgpool-II didn't find new primary node. Could you share the failover.sh? ===== 2019-01-04 11:30:42: pid 6456: DEBUG: do_query: extended:0 query:"SELECT pg_is_in_recovery()" 2019-01-04 11:30:42: pid 6456: DEBUG: verify_backend_node_status: there's no primary node 2019-01-04 11:30:42: pid 6456: LOG: find_primary_node: standby node is 0 ... 2019-01-04 11:30:45: pid 6456: LOG: failover: set new primary node: -1 2019-01-04 11:30:45: pid 6456: LOG: failover: set new master node: 0 ====== |
|
|
No response more than one month. Close issue. |
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2019-01-05 00:05 | jprieto | New Issue | |
| 2019-01-05 00:05 | jprieto | File Added: pgpool.conf.master-server | |
| 2019-01-05 00:05 | jprieto | File Added: pgpool.conf.slave-server | |
| 2019-01-05 00:05 | jprieto | File Added: pgpool.log.master-server | |
| 2019-01-05 00:05 | jprieto | File Added: pgpool.log.slave-server | |
| 2019-01-05 00:05 | jprieto | Tag Attached: logs | |
| 2019-01-05 00:07 | jprieto | Tag Detached: logs | |
| 2019-01-07 09:43 | pengbo | Note Added: 0002313 | |
| 2019-01-09 10:04 | pengbo | Assigned To | => pengbo |
| 2019-01-09 10:04 | pengbo | Status | new => feedback |
| 2019-02-18 14:31 | administrator | Note Added: 0002391 | |
| 2019-02-18 14:31 | administrator | Status | feedback => closed |
