>
> I've had trouble in the past with the vacuum command taking a long time
> to complete. This past vacuum went for 5 days before I CTRL-C'd it. I'm
> wondering if there is a bug.
>
> I have PostgreSQL 6.3.2 on a Red Hat 5.1 system (PII/400 MHz, 256 Meg
> RAM, 512 Meg Swap, 18 Gig Hard Ultra-wide SCSI Drive). The database
> consists of about 30 tables and is 1.2 Gig in total size.
>
> When I ran 'vacuum verbose analyze' the last time, the vacuum kept
> working for 5 days until I killed it. I had no other programs running on
> the system at the time. Also, I noticed that although 'top' showed that
> the vacuum was using 98% of the system resources, the harddrive activity
> was quiet during that 5 day period (as if nothing was actually
> happening).
That behaviour reminds me of a similar situation, where an
index of one of the user tables was corrupt, causing vacuum
to loop on that (all blocks cached so no HD activity).
Dropping/recreating the index in question solved that
problem.
But the difference this time is that it does not occur when
you manually vacuum all the user tables. So if it's the same
reason (corrupt index), this time it must be one of a system
catalog.
Can you dump/initdb/reload your database?
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#======================================== jwieck@debis.com (Jan Wieck) #