SpamAssassin timed out and was killed... box too slow?

mikea mikea at MIKEA.ATH.CX
Mon Jun 2 22:00:58 IST 2003


On Mon, Jun 02, 2003 at 01:53:03PM -0700, Chris W. Parker wrote:
> Kevin Spicer <mailto:kevins at BMRB.CO.UK> wrote:
>
> > > Actually I've got RBL checking turned off and my SA timeout is
> > > already set at 60s.
> >
> > > Any other ideas?
> >
> > Well, IIRC SA does some lookups anyway (even with rbls turned off),
> > not to mention any of the razor, pyzor, dcc checks you may or may not
> > be using.  Anyway its nothing to worry about if its just the
> > occasional message.
>
> I would agree except that I think it happens more than occasionally. Let me ask this question, even though it seems to time out frequently the consecutive failure count usually doesn't go past one. Here is an example, a few lines from my log. (i modified each line to try and shorten it as much as I could.) (These are all from today.)
>
> Jun 2 12:23:03 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 12:23:36 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 12:46:38 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 12:48:21 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 13:16:31 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 13:16:36 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 13:16:36 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 13:27:05 MS[nn]: SA timed out and was killed, consecutive failure 1 of 20
> Jun 2 13:29:03 MS[nn]: SA timed out and was killed, consecutive failure 2 of 20
> Jun 2 13:33:22 MS[nn]: SA timed out and was killed, consecutive failure 2 of 20
> Jun 2 13:35:12 MS[nn]: SA timed out and was killed, consecutive failure 2 of 20
> Jun 2 13:42:31 MS[nn]: SA timed out and was killed, consecutive failure 3 of 20
>
> Does the consecutive failure count get reset every few minutes or something? Otherwise I don't understand why the number stays at 1 so much. It looks like it is well on it's way to 20 but the last timeout only shows 3.
>
> It seems to me that if the count went to 20 all the time it would indicate that the box is indeed too slow.

The [nn] is the process ID for the process that is timing out. Every
time MS starts a new MS process, the timer restarts.

It would be nice if you would wrap your lines somewhere around 65 to 75 characters, possibly excepting quoted or copied lines such as maillog entries, so that they don't wind up looking like this, because not everyone can read terribly long lines with the same degree of ease, and indeed some people can't read them at all. OK? Thanks.

--
Mike Andrews
mikea at mikea.ath.cx
Tired old sysadmin since 1964



More information about the MailScanner mailing list