[Pgpool-general] Adding new slave node, need to restart pgpool-II?

[Guillaume Lelarge]

Le 01/04/2011 03:45, Sean Tegtmeyer a écrit :
> Hi all,
>     When adding in a new slave node to a running database cluster using
> pgpool-II w/Postgres9.0 SR, we would need to add a new *"**backend_hostname"
> *to the pgpool.conf, correct?  Then would it be necessary to restart
> pgpool-II as the pgpool-II manual suggests?  I ask because I thought I read
> somewhere else that it was not always necessary to restart pgpool-II when
> changes were made to the pgpool.conf config file.
> 

It's quite the same than PostgreSQL. Some changes don't require to
restart pgpool, and some require a restart. Unfortunately in your case,
adding a new backend needs a restart. Changing informations on one
backend doesn't require a restart.

> Is it also true that there is no other way to tell pgpool-II about a new
> node except via it's config file (some other way to avoid having to restart
> pgpool-II)?
> 

True.

> Also I read in the manual with regards to online recovery in master-slave
> mode with streaming replication that "To recover the primary node, you have
> to stop all DB nodes and pgpool-II, and then restore it from a backup."  Is
> this still true?
> 

What do you mean with "recover the primary node"?


-- 
Guillaume
 http://www.postgresql.fr
 http://dalibo.com