[pgpool-general: 4145] Re: manual switchover in pgpool 9.3

[Tatsuo Ishii]

> Hi,
> I haven't figured it out yet...
> I have 2 postgresql (9.3) nodes configured as master/slave (with streaming
> replication).
> I have configured pgpool as master/slave in order to perform failover in
> case of primary server goes down! everything is working great!
> 1. master server goes down.
> 2. failover script is being performed on the 2nd (slave) node. promote it
> into primary!
> 
> in order to recover the failed node (former primary):
> 3. pcp_recovery_node [params] -which  perform the recovery_1st_stage by
> performing pg_basebackup and creates recovery.conf file!!
> 
> *former primary is back as slave node!!!*
> 
> *my question is:*
> how can I perform manual switchover between the nodes (master into slave
> and slave into master) *without* performing a node recovery to the
> "switched off" primary. in other words , I don't want to perform a full
> base backup on the former primary node (its a large DB).
> 
> I have performed these steps(master/slave switchover) manually (without
> pgpool) with no pg_basebackup and its working great...

I think that's because no tranction commit nor checkpointing happened
during the operation. Othewise, base backup should be required.

Another to avoid the base backup is, using pg_rewind.

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