[pgpool-general: 8944] Re: pgpool 4.4.4: reading status file: 1 th backend is set to down status

[Tatsuo Ishii]

> Hi,
> 
> I've a cluster with 3 nodes. Every node runs bitnami/pgpool:4.4.4 as proxy and bitnami/postgresql-repmgr:15.4.0 as server. A PostgreSQL connection to all services (pg0, pg1, pg2, pgpool0, pgpool1, pgpool2) can be established.
> In the QA system I see that pgpool of node 1 is not running properly. It is always in state "starting" and never "healthy". Therefore orchestration is terminating and restarting the container.
> Having a look at the log of pgpool1 and comparing it with the other pgpool instances there is one difference:
> main pid 1: LOG:  reading status file: 1 th backend is set to down status
> 
> Therefore my questions:
> 
>   1.  Is there a file which buffers pg states?

If you mean "pg_status" column in show pool_nodes command, no. It is
obtained from PostgreSQL on the fly when show pool_nodes command gets
executed.

>   2.  How did the system get into this state?

I am not familiar with bitnami pgpool nor repmgr. So what I can do is
answer from the point of pgpool view. It was caused by either failover
triggered by health check (pgpool detects error / shutdown of
PostgreSQL), or pcp_detach_node gets executed. I cannot tell either
unless looking into pgpool log and pgpool.conf

Best reagards,
--
Tatsuo Ishii
SRA OSS LLC
English: http://www.sraoss.co.jp/index_en/
Japanese:http://www.sraoss.co.jp