Re: Need help with Dead Lock - reg. - Mailing list pgsql-general

From Tom Lane
Subject Re: Need help with Dead Lock - reg.
Date
Msg-id 9305.1087871055@sss.pgh.pa.us
Whole thread Raw
In response to Re: Need help with Dead Lock - reg.  (Mike G <mike@thegodshalls.com>)
List pgsql-general
Mike G <mike@thegodshalls.com> writes:
> On Sat, 2004-06-19 at 13:12, Shanmugasundaram Doraisamy wrote:
>> Is there a way to increase the time of Dead Lock?
>> The default is set to 1000 msec.

> I would consider the default probably an optimum value.  If you are
> having to increase it to keep your applications working then there are
> other issues that need to be addressed.

AFAIK, neither increasing nor decreasing that timeout will have the
slightest effect on whether your apps get deadlock failures.  All it
does is determine how quickly an existing deadlock is detected.
A shorter timeout means you wait less long to be told of a deadlock,
but it also causes more cycles to be spent checking for deadlocks
that may not exist.  If you've got apps that are designed not to ever
deadlock, then you might want to kick the timeout up a lot to ensure
you don't waste time on deadlock checks even under heavy load.

So I'm wondering what the OP's problem *really* is; whatever it is,
I doubt that fooling with deadlock_timeout will solve it.

            regards, tom lane

pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Problem while OS-Update
Next
From: Dennis Gearon
Date:
Subject: who can run utility - pg_controldata ?