[pgpool-general: 1772] Re: Pgpool with backends on the same box problem
stelios.limnaios at mekon.com
Thu May 23 16:30:09 JST 2013
Although I didn't hit that FAQ page on my searching, it seems that the answer is not there.
I tried to manually run the pcp_* commands on the console and they ran ok (got AuthenticationOK and CommandCompleted).
This suggests that there is a permissions issue that prevents pgpooladmin from running the pcp_* programs properly.
The next to do was to disable SELinux, and that fixed the problem!
Pgpooladmin was displaying the information on nodes and its status fine.
So, the question is different now, to my opinion.
What SELinux permissions must we grant to the pcp_* programs?
Again, many thanks for your help and time.
From: Tatsuo Ishii [mailto:ishii at postgresql.org]
Sent: 22 May 2013 16:54
To: Stelios Limnaios
Cc: pgpool-general at pgpool.net
Subject: Re: [pgpool-general: 1767] Pgpool with backends on the same box problem
> I would like to ask for some help on a strange issue we have.
> We have successfully installed pgpool in the past, on one of our servers with 2 back-ends each one on a separate server.
> Everything works ok on this cluster.
> Now, we're trying to install a similar cluster (pgpool and 2 backends) on a separate server for testing purposes.
> We want to install everything on one box.
> The problem we have is that pgpooladmin shows the nodes as down although the databases are running and the ports are open to the network. Also, when we start pgpool from the pgpooladmin, we get e1002 - pcp_node_count command error occurred.
> The strange is that if we use a backend that is located on a different server pgpooladmin shows fine its status.
> What are we missing? Is it something related to permissions? Any direction will be much appreciated.
Have you checked pgooolAdmin FAQ?
SRA OSS, Inc. Japan
More information about the pgpool-general