[pgpool-general: 6958] Re: Fwd: pcp_attach_node downtime

[Tatsuo Ishii]

> At times, the "standby" node might go down for whatever reason. When I
> issue pcp_attach_node command to start sending queries to it - the
> whole pgpool interface seems to be disrupted:
> 
> 
> - clients see disruption in queries
> 
> - no new queries are going through
> 
> - new connections to pgpool fail
> 
> Normal functionality is restored after about 2-5 seconds.
> 
> 
> Is this an expected behavior

Yes. Pgpool-II needs to prohibit access from clients while failback
operation is in progress.

> or is there a more graceful way to
> reattach a node in pgpool?

Currently no. However I am thinking about improving this. For example,
if failback is in progress, wait until failback ends. This will be a
slow down of new queries from client's point of view.

Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp