Re: vacuum is failing

I've got a server running PostgreSQL-7.2.3 on RHAS-2.1. The version
of Postgresql is *not* my choice, I would have gone with something
much more current, but i digress.

Running 'vacuumdb -a -z -v -f' has started failing:

NOTICE: Index pg_toast_218644531_idx: Pages 1; Tuples 0.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
NOTICE: Analyzing doc2_review_comments
NOTICE: --Relation groups--
NOTICE: Pages 28: Changed 0, reaped 2, Empty 0, New 0; Tup 331: Vac
2, Keep/VTL 0/0, UnUsed 1, MinLen 160, MaxLen 1988; Re-using:
Free/Avail. Space 13360/12448; EndEmpty/Avail. Pages 0/17.
CPU 0.01s/0.00u sec elapsed 0.07 sec.
NOTICE: Index grps: Pages 2; Tuples 331: Deleted 2.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
NOTICE: Index groups_public: Pages 2; Tuples 331: Deleted 2.
CPU 0.00s/0.00u sec elapsed 0.02 sec.
NOTICE: Index groups_status: Pages 2; Tuples 331: Deleted 2.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
NOTICE: Index groups_type: Pages 2; Tuples 331: Deleted 2.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
NOTICE: Index group_unix_uniq: Pages 4; Tuples 331: Deleted 2.
CPU 0.00s/0.00u sec elapsed 0.00 sec.
ERROR: No one parent tuple was found

Anyone have any suggestions on how to correct this? This DB serves a
web based PHP application, and right now there isn't any end user
noticable problems, and i'd like to keep it that way :)

thanks in advance!

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L. Friedman netllama(at)gmail(dot)com
LlamaLand http://netllama.linux-sxs.org

Lonni J Friedman <netllama(at)gmail(dot)com> writes:
> ERROR: No one parent tuple was found

> Anyone have any suggestions on how to correct this?

Shut down whatever long-running transaction is hanging around in the
background. You wouldn't be getting this error if vacuum weren't
trying to move a chain of updated tuples, and that can only happen
if there's an old transaction someplace.

Of course, the real fix is to use a newer Postgres. I note this in
the 7.2.4 CVS logs:

2003-01-26 18:09 tgl

* src/backend/access/heap/heapam.c (REL7_2_STABLE): Back-patch
fixes to ensure t_ctid always has correct value (prevents some
instances of 'No one parent tuple' VACUUM error, and perhaps worse
things).

regards, tom lane

On Tue, 04 Jan 2005 18:30:43 -0500, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Lonni J Friedman <netllama(at)gmail(dot)com> writes:
> > ERROR: No one parent tuple was found
>
> > Anyone have any suggestions on how to correct this?
>
> Shut down whatever long-running transaction is hanging around in the
> background. You wouldn't be getting this error if vacuum weren't
> trying to move a chain of updated tuples, and that can only happen
> if there's an old transaction someplace.
>
> Of course, the real fix is to use a newer Postgres. I note this in
> the 7.2.4 CVS logs:
>
> 2003-01-26 18:09 tgl
>
> * src/backend/access/heap/heapam.c (REL7_2_STABLE): Back-patch
> fixes to ensure t_ctid always has correct value (prevents some
> instances of 'No one parent tuple' VACUUM error, and perhaps worse
> things).

Thanks Tom. I should have googled first, cause this seems to have
come up alot before 7.2.4 was released. I'll schedule some down time
and see if running vacuum fixes this (temporarily). Getting
postgresql upgraded might require an act of $DEITY *sigh*

--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L. Friedman netllama(at)gmail(dot)com
LlamaLand http://netllama.linux-sxs.org