[Pgpool-hackers] Follow master patch modified

[Gilles Darold]

Le 06/04/2011 04:25, Toshihiro Kitagawa a écrit :
> Hi Gilles,
>
> Thank you for comment.
>
> Could you test the patch?
> Does it resolve the following problem?
>
>   Subject: [Pgpool-hackers] Major problem in find_primary_node()
>   Date: Wed, 19 Jan 2011 19:43:16 +0100
>
> Regards,
>
Hi Toshihiro,

I've just tested this patch, it can solve the problem on some part but
it raise a main problem in pool_auth.c:

     pool_read_kind: kind does not match between master(0) slot[0] (52)

Everything is working well if the failover destination master is node 0
but with other pgpool is complaining about this error. This is purely
pgpool related as all PostgreSQL nodes are ok, the master is replicated
on other slaves without any problems. Maybe the reason is related to the
use of IS_MASTER_NODE_ID instead of PRIMARY_NODE_ID, but I've not
investigate more. I don't know what pool_read_kind() is doing if you can
explain me, I will take the few coming days to work on this patch.

Regards

-- 
Gilles Darold
http://dalibo.com - http://dalibo.org