Re: BUG #3748: Postmaster Service

From: Andrew Sullivan <ajs(at)crankycanuck(dot)ca>
To: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #3748: Postmaster Service
Date: 2007-11-15 13:12:46
Message-ID: 20071115131246.GJ19518@crankycanuck.ca
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

A couple things:

On Thu, Nov 15, 2007 at 12:49:40PM +0000, Sudarshan V Tantry wrote:
> PostgreSQL version: 7.8.2

There is no PostgreSQL version numbered that.

> We are supporting a messaging system in which there are two application
> servers with cluster services.Cluster service does the job of shifting the
> Application and Database services from one server to another if one of the
> Application servers goes down.

So you have two application servers and two database servers? Two servers
with the application and database on each? I don't really understand.
Also,

> services,by manually stopping cluster service on one server using the
> command "service clumanager stop" the Application and Database services get
> shifted to another server,but Application services would be down due to the
> fact that it cannot connect to the Database services since the postmaster
> service is down.We check the status of Postmaster service with the command

In what sense do they get shifted to another server, if the postmaster
doesn't start up?

> Also postmaster services do not start even when the server is restarted.

Is there a start up script for postgres?

>
> DBDATA is the database service that is mounted on the application
> server.

I don't understand this sentence either. But. . .

> That is getting mounted but postmaster would be found stopped.

. . .mounting a directory does not entail starting a database server. These
are separate problems. You have to have the PostgreSQL data area mounted,
and _then_ start PostgreSQL. Perhaps your cluster thingy isn't doing it in
the right order.

In any case,

> Please let me a solution for this problem and also tell me why this problem
> occurs.

this doesn't seem to be a bug with PostgreSQL at all. If anything, it has
something to do with this cluster thing you're using. That's in general a
dangerous thing to do, and requires plenty of care. See the high
availability parts of the manual to read general discussions of this, but
if you're having specific problems, I think you need to talk to the people
who make the cluster software.

A

--
Andrew Sullivan
Old sigs will return after re-constitution of blue smoke

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Christian Ullrich 2007-11-15 18:22:03 BUG #3750: Invalid frontend message type 112
Previous Message Sudarshan V Tantry 2007-11-15 12:49:40 BUG #3748: Postmaster Service