Re: Postgres startup problem - Mailing list pgsql-hackers

From Nataraj
Subject Re: Postgres startup problem
Date
Msg-id 39B00D61.DEEA8740@cdxc.com
Whole thread Raw
In response to Re: Postgres startup problem  (John McKown <jmckown@prodigy.net>)
List pgsql-hackers
Thanks John.
I will try this when I reconstruct the server from the backup. Just to keep the system
going, I have temporarily reconstructed the entire server from an old Ghost backup. I
will bringback the image of the server which had the postgres startup problem because
I need to recover all the data from the database.

Thanks,
Nataraj

John McKown wrote:

> The problem is that the "socket" is actually a file. On my system, this
> file is "/tmp/.s.PGSQL.5432". Logon as root. Take down PostgreSQL. Then
> "rm" that file. PostgreSQL should then start up OK.
>
> On Fri, 1 Sep 2000, Nataraj wrote:
>
> > Hi all,
> > I have a problem starting the postgres database.
> > I am using postgres 7.0.2 on a Redhat Linux 6.2. I had to shutdown my computer
> > since it was hanging. I have the postgres start command added to the bootup
> > process. I have rebooted the Linux machine several times and it was starting
> > postgres correctly. But this time it was saying "Postmaster could not connect to
> > unix socket 5432" . I checked if any other instance of Postmaster was running by
> > checking the process ids. There was no second instance running. Also, I tried to
> > stop and restart postgres using
> > "/etc/rc.d/init.d/postgres stop" and "/etc/rc.d/init.d/postgres start" commands.
> > It was saying
> > "Starting Postres [ ]". If Postgres was really started it will show the process
> > id within the square brackets. But this time it did not show the process id. If I
> > try to connect to the database using psql it gives the error message "Postgres
> > could not be connected to socket 5432".
> >
> > Please let me know if there is a different way of starting  postgres.
> > Thanks,
> > Nataraj
> >



pgsql-hackers by date:

Previous
From: John McKown
Date:
Subject: Re: Postgres startup problem
Next
From: Tom Lane
Date:
Subject: A fine point about OUTER JOIN semantics