[Pgpool-general] Database nodes mismatch after failback

Marcos Davi Reis davi at movamaps.com
Mon Dec 14 16:29:25 UTC 2009


Tomasz,
The DB wasn't stopped! now the recovery works properly! tks a lot!!


i'm still having messages about difference between nodes, and sometimes the
node is degenerated, sometimes not..
the degeneration doesn't work when the difference is detected in a simple
update query, but works when a stored procedure who changes data, is called.

Example:

Dec 14 14:20:39 pgsql1 pgpool: 2009-12-14 14:20:39 LOG:   pid 7474:
statement: /*REPLICATION*/ select mova_rcv_tracker_data(332617,'2009-12-14
16:20:29'::timestamp without time zone, -2.53214, -44.1992, 0 , 2, 0,
1::boolean, 0::boolean, 0::boolean, 0::boolean, 0::boolean, 0::boolean,
0::boolean, 0::boolean, 0::boolean, 1::boolean,1::boolean, 1524742 6, 509,
13, 0, 0, 0, 0, 24069, 0::boolean)


and i can't find what is causing the difference! I supose it can be caused
by stored procedures who changes data (insert, update and delete).
how does pgpool work with this statements?




Att,
Marcos Davi Reis
Mova


On Mon, Dec 14, 2009 at 12:22 PM, Tomasz Chmielewski <mangoo at wpkg.org>wrote:

> On 14.12.2009 16:09, Marcos Davi Reis wrote:
>
> (...)
>
>
>  Dec 14 12:00:05 pgsql1 pgpool[5503]: Executing pg_stop_backup
>> Dec 14 12:02:05 pgsql1 pgpool[5529]: Executing pgpool-recovery-pitr as
>> user postgres
>> Dec 14 12:02:05 pgsql1 pgpool[5530]: Executing pg_switch_xlog
>> Dec 14 12:02:05 pgsql1 pgpool[5539]: Executing pgpool_remote_start as
>> user postgres
>> Dec 14 12:02:05 pgsql1 pgpool[5540]: Starting remote PostgreSQL server
>>
>>
>> ps. I used the Jaume's article (at linuxsilo) as reference to implement
>> pgpool...
>>
>
> Do you stop Postgres on remote node _prior_ to rsync?
>
> It is not (was not?) clearly explained, nor outlined in example scripts.
>
> See this thread from yesterday:
>
>    "torture test" for pgpool-ii?
>
>
> --
> Tomasz Chmielewski
> http://wpkg.org
>



-- 
Marcos Davi Reis
Mova
www.movaomundo.com
+55 21 3553-1511
+55 21 9923-8319
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pgfoundry.org/pipermail/pgpool-general/attachments/20091214/393a9831/attachment-0001.html>


More information about the Pgpool-general mailing list