[pgpool-general: 9294] Re: Failover or Failback event detected

[Tatsuo Ishii]

> Hello,
> 
> Along with other services running in docker swarm architecture I have a
> database server of 3 PostgreSQL nodes deployed in Patroni + ETCD HA
> Solution and a connection pooling method, PgPool-II.
> 
> PostgreSQL Version: 13.8
> PgPool Version: 4.2.5
> HA Cluster-Management of Database: Patroni + ETCD (1 Primary + 2 Standby)
> 
> ** Note: Load Balancing is disabled. All the read-write operations are
> routed only to Primary Node.
> 
> *Issue:* Frequent logs related to failover or failback event detected.
> Pgpool-II Logs:
> 
>> 2024-12-19 18:09:45: pid 776: LOG:  failover or failback event detected
>> 2024-12-19 18:09:45: pid 776: DETAIL:  restarting myself
> 
> 
> I also tracked the database logs, but didn't find any logs relevant to the
> issue.
> Database Logs:
> 
>> 2024-12-19 18:09:45 UTC [81078]: [67646169.13cb6-1]
>> [unknown]@[unknown],app=[unknown] [00000] LOG:  connection received:
>> host=10.10.1.122 port=42172
>> 2024-12-19 18:09:45 UTC [81078]: [67646169.13cb6-2] tsdb at tsdb,app=[unknown]
>> [00000] LOG:  connection authorized: user=tsdb database=tsdb
> 
> 
> As per my understanding, these pgpool-II logs occur when Pgpool-II is not
> able to communicate with PostgreSQL nodes. Which happens probably when
> Database is not running or service is restarted.
> But the database service seems to be running normally.

Yes, but in addition to this, communication error between pgpool and
PostgreSQL is treated as PostgreSQL down event too.

> Please help on this issue. Am I missing something to check or there can be
> any other issue?

Do you turn off failover_on_backend_error?

Can you share pgpool.conf?  (you can make it blank sensitive
information such as password and IP address).

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