[Pgpool-general] re-attaching the failed backend

[Nenad Bulatović]

Hi,

I am using pgpool-II in raw mode to support failover between two backends.
master/slave, load_balancing, replication and connection cache are all set
to off. Databases are replicated using streaming replication, however we
just need simple failover between databases as database failover is to be
initiated manually.

failover_on_backend_error is  set to on and both backends are configured
with default flag ALLOW_TO_FAILOVER set.

Once backend0 (master) is down, pgpool failovers without any problems to
standby, leaving backend0 in status 3. After master is brought back again,
backend0 is not automatically reattached and it is impossible to re-attach
it manually using pcp_attach_node command.

My questions:

- does raw mode support automatic backend re-attach at all?

- why is it not possible to reattach node manually using pcp_attach node
command?

Attached is the configuration file

Best Regards,
Nenad
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://pgfoundry.org/pipermail/pgpool-general/attachments/20111109/57c8dc29/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pgpool.conf
Type: application/octet-stream
Size: 1703 bytes
Desc: not available
URL: <http://pgfoundry.org/pipermail/pgpool-general/attachments/20111109/57c8dc29/attachment.obj>