Thread: 8.0B2 checkpoint error on start up

Upgrading 8.0b1 to 8.0b2 in place.
Did I have to do a dump/restore?
Or is this a known problem?

--elein

$ pg_ctl start -D $PGDATA
postmaster starting
/local/pgsql80/data[postgres:cookie]
$ LOG:  database system was shut down at 2004-09-14 13:40:43 PDT
LOG:  invalid magic number D05B in log file 0, segment 34, offset 7495680
LOG:  invalid primary checkpoint record
LOG:  invalid magic number D05B in log file 0, segment 34, offset 7438336
LOG:  invalid secondary checkpoint record
PANIC:  could not locate a valid checkpoint record
LOG:  startup process (PID 4574) was terminated by signal 6
LOG:  aborting startup due to startup process failure

Quick followup...The 8.0 db was only written to by slon
processes.  This problem was replicated by someone else
as well.

The stomp around was to initdb in 8.0b2 and reset up
slony :-(

--elein

On Tue, Sep 14, 2004 at 01:48:54PM -0700, elein wrote:
> Upgrading 8.0b1 to 8.0b2 in place.
> Did I have to do a dump/restore?
> Or is this a known problem?
>
> --elein
>
> $ pg_ctl start -D $PGDATA
> postmaster starting
> /local/pgsql80/data[postgres:cookie]
> $ LOG:  database system was shut down at 2004-09-14 13:40:43 PDT
> LOG:  invalid magic number D05B in log file 0, segment 34, offset 7495680
> LOG:  invalid primary checkpoint record
> LOG:  invalid magic number D05B in log file 0, segment 34, offset 7438336
> LOG:  invalid secondary checkpoint record
> PANIC:  could not locate a valid checkpoint record
> LOG:  startup process (PID 4574) was terminated by signal 6
> LOG:  aborting startup due to startup process failure
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org

elein <elein@varlena.com> writes:
> Upgrading 8.0b1 to 8.0b2 in place.
> Did I have to do a dump/restore?

No, but I think you need a pg_resetxlog.

            regards, tom lane