[Pgpool-general] Database redundancy recommended approach

[Tatsuo Ishii]

>>From your point of view, what is the recommended approach to have the
> following scenario working without human intervention?
> 
>  
> 
> *         2 DB Servers (DB0 and DB1)
> 
> *         1 pgpool instance (in a thrid server)
> 
> *         Streaming replication from DB0 (master) to DB1 (slave)
> 
> *         If DB0 fails, DB1 must become primary.

Easy.

> *         As soon as DB0 is ready again, make it slave of DB1.

With 3.0.2(or CVS HEAD) it's possible if you write a program(outside
pgpool world) which triggers online recovery when DB0 is ready again.

3.0.1 or before does not handle the case when DB0 is standby and DB1
is primary.

> *         With DB1 as primary, if it fails, DB0 must become primary
> (again).

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