View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000662 | pgpool-HA | Bug | public | 2020-11-29 05:52 | 2020-11-30 12:42 |
| Reporter | jeffty_w | Assigned To | |||
| Priority | high | Severity | major | Reproducibility | always |
| Status | closed | Resolution | open | ||
| Platform | Linux | OS | RHEL | OS Version | 7.7 |
| Summary | 0000662: PgPool with WatchDog Cluster doesn't work after reboot all the machines | ||||
| Description | Hi there, I installed Pgpool + WatchDog as the example https://www.pgpool.net/docs/latest/en/html/example-cluster.html described on 3 machines. Today they are powered off by accident. After power on them I found that db2 and db3 are in 'down' status. Is there a recommended way to auto start all of the nodes even after reboot? RPMs: postgresql: 12.3 pgpool-II-pg12: 4.1.2 Machines: jeffty-db11, jeffty-db2, jeffty-db3 VIP: 192.168.1.208 Services Configured during the installation: pgpool service is not enabled, systemctl enable postgresql-12 on every machine. cat /usr/lib/systemd/system/postgresql-12.service ExecStartPre=/usr/pgsql-12/bin/postgresql-12-check-db-dir ${PGDATA} ExecStart=/usr/pgsql-12/bin/postgres -D ${PGDATA} And here is how I initiated the recovery nodes on db1 several months ago: pcp_recovery_node -h 192.168.1.208 -p 9898 -U pgpool -n 1 pcp_recovery_node -h 192.168.1.208 -p 9898 -U pgpool -n 2 Now after power on those machines, only the db1's postgresql service is active. I started the pgpool service manually one by one and check the watchdog info on db1: -bash-4.2$ pcp_watchdog_info -h 192.168.1.208 -p 9898 -U pgpool 3 YES jeffty-db1:9999 Linux jeffty-db1 jeffty-db1 jeffty-db1:9999 Linux jeffty-db1 jeffty-db1 9999 9000 4 MASTER jeffty-db2:9999 Linux jeffty-db2 jeffty-db2 9999 9000 7 STANDBY jeffty-db3:9999 Linux jeffty-db3 jeffty-db3 9999 9000 7 STANDBY Looks good. and I tried -bash-4.2$ psql -h 192.168.1.208 -p 9999 -U pgpool postgres -c "show pool_nodes" node_id | hostname | port | status | lb_weight | role | select_cnt | load_balance_node | replication_delay | replication_state | replication_sync_state | last_status_change ---------+--------------------+------+--------+-----------+---------+------------+-------------------+-------------------+-------------------+------------------------+--------------------- 0 | jeffty-db1 | 5432 | up | 0.333333 | primary | 247 | true | 0 | | | 2020-11-28 14:04:24 1 | jeffty-db2 | 5432 | down | 0.333333 | standby | 0 | false | 0 | | | 2020-11-28 14:01:26 2 | jeffty-db3 | 5432 | down | 0.333333 | standby | 0 | false | 0 | | | 2020-11-28 14:01:28 (3 rows) db2 & db3 are in 'down' status. Then I login them and tried: db2#su - postgres -bash-4.2$ psql -U postgres psql: error: could not connect to server: could not connect to server: No such file or directory Is the server running locally and accepting connections on Unix domain socket "/var/run/postgresql/.s.PGSQL.5432"? So I have to start postgresql manually as /usr/pgsql-12/bin/pg_ctl -D /var/lib/pgsql/12/data/ start Now I can use psql -U postgres to login db2, but it seems that they are not in replicated since I create a table on db2 and cannot find the new table in db1. And finally after rebooting: 1. If I don't manually start postgresql service on db2 & db3, then it turns into only one node cluster, db1 acts as the only database instance. 2. If I manually start postgresql service on db2 & db3, it should still be one node instance since changes on db2 won't be synchronized to db1. What steps I missed during this process? Any suggestion or guidance would be appreciated. Best Regards, Jeffty | ||||
| Steps To Reproduce | 1. Install pgpool 4.1.2 and postgresql 12.3 on 3 machines as https://www.pgpool.net/docs/latest/en/html/example-cluster.html described. 2. enable the postgresql-12 services. 3. don't enable pgpool services. 4. reboot all of the machines. | ||||
| Tags | No tags attached. | ||||
|
|
This is not a issue related "pgpool-HA". Could you close this issue and create a new one with project "Pgpool-II"? |
|
|
Sorry for that. New ticket: https://www.pgpool.net/mantisbt/view.php?id=664 |
|
|
Close issue |
