[Pgpool-general] online recovery of primary is possible or not?

[Tatsuo Ishii]

> I've read here:
> http://pgpool.projects.postgresql.org/contrib_docs/simple_sr_setting/index.html
> that online recovery with streaming replication is possible.
> But in the official doc it is said that only with standby it's possible to
> perform online recovery
> (http://pgpool.projects.postgresql.org/pgpool-II/doc/pgpool-en.html#master_slave_mode)...who
> is right?

Both are correct. In the former, here is the supposed scenario:

1) primary goes down
2) standby takes over and becomes primary
3) By using online recovery, the downed former primary could be online
   as a new standby.

> I've tried to configure pgpool to perform an online recovery anyway, as
> described in the first link, but without success. Unlike that tutorial I have
> three separated machines which are primary, standby and pgpool.
> 
> Everything works fine but at the very last step, postgres on the ex-primary (now
> standby) says:
> LOG:  entering standby mode
> LOG:  streaming replication successfully connected to primary
> LOG:  unexpected timeline ID 24 in log file 3, segment 60, offset 0
> FATAL:  terminating walreceiver process due to administrator command
> LOG:  unexpected timeline ID 24 in log file 3, segment 60, offset 0

As far as I can tell, it seems the base backup from primary to standby
was not done properly. What is your baseback.sh script exactly?
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp