Martijn van Oosterhout wrote:
On Mon, Aug 27, 2007 at 06:37:17PM +0200, Kamil Srot wrote:
I don't say, it's gone by itself, I'm asking for help debuging this
situation and hopefully find a solution. For the first time it happened,
it had the same symptoms - this specific table was missing and
transaction counter was wrapped-around. Now it seems to be the same but
the transaction counters are far below the critical value...
You don't say explicitly, but when you say "vacuum all tables" you
didn't specify any tables names right? You just typed "vacuum". On each
database...
Correct...the script does echo "vacuum full;" | $PGDIR/bin/psql -U postgres $db for each database...
Hope it's correct?
As for debugging, maybe something like:
select xmin, age(xmin) from pg_class;
Just to check the wraparound issue...
Maybe it's it... I did check the wraparound with:
SELECT datname, age(datfrozenxid) FROM pg_database;
In your SQL I see several rows with too high numbers!
They are all "internal" like pg_toast_618854, views and also some application level indices etc.
So it seems it's not enough to vacuum all DB's (templateX inclusive)
Can you please advice how to correctly manually vacuum the pgsql?
Thank you!
Regards,
--
Kamil