Thread: Error when trying to do a recovery

Error when trying to do a recovery

From
Warren Little
Date:
Hello,
I trying to do a restore using from the online backups.  I have restored
the data backup (untarred) and removed all files from the pg_xlog
directory.  I also deleted and recreated the archive_status directory
under pg_xlog.  I created the recovery.conf file with only the
restore_command in it.
When I start postgres (pg_ctl -l logfile start). The following text is
all that exists in the logfile:

<2005-03-01 14:21:43 MST>FATAL:  incorrect checksum in control file


What control file is it referring to?  Is there some way to generate
more verbose logging?


thanks

Warren Little
Meridias Capital


Re: Error when trying to do a recovery

From
Tom Lane
Date:
Warren Little <warren.little@meridiascapital.com> writes:
> When I start postgres (pg_ctl -l logfile start). The following text is
> all that exists in the logfile:

> <2005-03-01 14:21:43 MST>FATAL:  incorrect checksum in control file

> What control file is it referring to?

$PGDATA/global/pg_control.  I think this must mean that you are trying
to start a postmaster that is fundamentally incompatible with the old
postmaster --- enough so that it thinks the pg_control struct should be
a different size than the old postmaster thought.

            regards, tom lane

Re: Error when trying to do a recovery

From
Warren Little
Date:
Tom,
Thanks for your response.  I found an old thread you had responded to a while back
regarding 32 vs 64 bit architectures (which I have) and compatibility issues.
So until I upgrade my backup/test box looks like I'm stuck using pg_dump :(

Thanks


On Wed, 2005-03-02 at 01:22 -0500, Tom Lane wrote:
Warren Little <warren.little@meridiascapital.com> writes:
> When I start postgres (pg_ctl -l logfile start). The following text is
> all that exists in the logfile:

> <2005-03-01 14:21:43 MST>FATAL:  incorrect checksum in control file

> What control file is it referring to?

$PGDATA/global/pg_control.  I think this must mean that you are trying
to start a postmaster that is fundamentally incompatible with the old
postmaster --- enough so that it thinks the pg_control struct should be
a different size than the old postmaster thought.

			regards, tom lane