pgsqld_monitor_0 on <hostname> 'not configured' about paf HOT 5 CLOSED

Hello @Netbulae,

There's various ways PAF could refuse to deal with your setup. Usually it explains what is wrong in log files so you can fix the issue on your side. See here for the various checks PAF is doing in 2.0: https://github.com/dalibo/PAF/blob/v2.0.0/script/pgsqlms#L984

However, maybe you felt in the only known incompatibility between PAF and PostgreSQL 9.6: how did you set your wal_level parameter? We accepted a pull request few days ago about wal_level = replica, see #47.

This fix will be included in the next release.

Regards,

from paf.

Any news on this issue? Should I close it?

from paf.

Hi, still the same issue. Trying to find anything in the log. What I can see is:

Nov 21 15:57:29 [3494] database1.test.nu pengine: warning: unpack_rsc_op_failure: Processing failed op stop for pgsqld:0 on database1: not configured (6)
Nov 21 15:57:29 [3494] database1.test.nu pengine: error: unpack_rsc_op: Preventing psql-ha from re-starting anywhere: operation stop failed 'not configured' (6)
Nov 21 15:57:29 [3494] database1.test.nu pengine: warning: unpack_rsc_op_failure: Processing failed op stop for pgsqld:0 on database1: not configured (6)
Nov 21 15:57:29 [3494] database1.test.nu pengine: error: unpack_rsc_op: Preventing psql-ha from re-starting anywhere: operation stop failed 'not configured' (6)

For now, I have both slave servers set to hot standby:
wal_level = hot_standby

from paf.

Hi,

Even if you set wal_level = hot_standby in postrgesql.conf, this parameter is set to replica internally.

The following command expose this incompatibility with current PAF 2.0:

$ psql -c "show wal_level"
 wal_level 
-----------
 replica

I'm afraid you either need to use current HEAD code, patch pgsqlms yourself using #47 or wait for release 2.1.

from paf.

I'm closing this issue as this has been fixed in #47 and will be included in next release.

Thank you for reporting. Do not hesitate to ask more questions if needed.

from paf.