Re: Idle in transaction ????

Meant to send this to the list......

---------------------- Forwarded by Wesley Sheldahl/Lex/Lexmark on 11/15/2001
02:37 PM ---------------------------

Wesley Sheldahl
11/15/2001 11:04 AM

To: Martijn van Oosterhout <kleptog%svana(dot)org(at)interlock(dot)lexmark(dot)com>
cc:

Subject: Re: [GENERAL] Idle in transaction ???? (Document link: Wesley
Sheldahl)

If vacuum in 7.2 skips tables it can't lock, I would hope there would at least
be something logged to that effect so problems like this can be found and
resolved. Otherwise, it seems that a table might go for weeks without being
successfully vacuumed if some process has it perpetually locked.

Wes Sheldahl

Martijn van Oosterhout <kleptog%svana(dot)org(at)interlock(dot)lexmark(dot)com> on 11/15/2001
09:45:21 AM

Please respond to Martijn van Oosterhout
<kleptog%svana(dot)org(at)interlock(dot)lexmark(dot)com>

To: Steve Brett <steve(dot)brett%e-mis(dot)com(at)interlock(dot)lexmark(dot)com>
cc: pgsql-general%postgresql(dot)org(at)interlock(dot)lexmark(dot)com (bcc: Wesley
Sheldahl/Lex/Lexmark)
Subject: Re: [GENERAL] Idle in transaction ????

On Wed, Nov 14, 2001 at 03:21:36PM -0000, Steve Brett wrote:
[snip]
> it's the 'idle in transact' that i'm puzzled about and i'm assuming it's
> idel within a transaction block ...

You've got something with an active transaction somewhere and VACUUM is
waiting for it to finish. Find whichever process is guilty and fix it to
commit the transaction.

I think the new vacuum in 7.2 is less picky about locks and will simply skip
tables it can't get a lock on.
--
Martijn van Oosterhout <kleptog(at)svana(dot)org>
http://svana.org/kleptog/