[Pgpool-general] PgPool I/II Advice Required

[Mikko Partio]

Bruce McAlister wrote:
>
> Hi All,
>
> We have a requirement where our services need to be available 24/7 and 
> in case of a failure the downtime needs to be minimal. We have been 
> directed to use PGPool for the failover/connection pooling features. 
> My question is, which version of PGPool should we use to test the 
> software, PGPool I or PGPool II.
>
> We currently have the following scenario:
>
> HOST1:
>
> PostgreSQL 8.2.3 Live Production
>
> HOST2:
>
> PostgreSQL 8.2.3 Recovery Mode (Standby Server)
>
> HOST2 is at most 60 seconds out of sync with HOST1 using PostgreSQL 
> WAL shipping.
>
> What we would like to do is put PGPool in the front with HOST1 defined 
> as the primary node and HOST2 as the slave node. How would we go about 
> ensuring that PGPool detects a failure on HOST1 then automatically 
> starts to use HOST2. Remember though that HOST2 is not available for 
> health checks while it is in recovery mode, so I assume that PGPool 
> will run in degen mode. However, when the database on HOST2 does 
> become available will PGPool automatically start using it, or is it a 
> manual process to get PgPool to fail over it’s connections to HOST2 
> after it comes out of standby/recovery mode?
>
> Any and all help/tips/suggestions/comments/criticisms are most welcome.
>
> Thanks
>
> Bruce
>
I would also like to know if pgpool (especially -II) works well with wal 
shipping.

Regards

MP