Too many early morning pages

Gary Mills mills@cc.UManitoba.CA
Mon Feb 16 04:12:55 UTC 2004

On Sun, Feb 15, 2004 at 07:52:30AM -0700, Vernon Schryver wrote:
> > From: Gary Mills 
> > I had another incident this morning where sendmail was rejecting
> > connections.  This time, I was able to collect some better information.
> > The culprit seems to be `dbclean'.  
> > Feb 14 06:06:14 electra dccd[23006]: [ID 968768 mail.notice] 837646 free hash entries among 16760832 total; starting `dbclean -DPq -i 1032`
> That >95% hash table load is a reason to install 1.2.32 on electra
> as soon as possible.

Yes, it was happening once, and sometimes twice, a day.  I upgraded
this morning.

> I cannot imagine how dbclean could cause dccm to have lots of idle TCP
> sockets.  No matter what craziness dbclean might have been suffering, 
> the worst it might have done to dccm is make the system very slow.

Yes, that's probably what happened.  The background level of e-mail
from spam and viruses is quite high these days.  If anything slows
down the machine, it only takes a few minutes for sendmail to begin
hitting limits.  If the condition persists, sendmail begins to report
milter errors.

> > There is a second DCC server that was running normally at the time.
> > Why didn't `dccm' just start using it while `dbclean' was running on
> > the first one?
> Does /var/dcc/map allow that, as indicated by `cdcc info`? 
> What happens if you shut down dccd with `cdcc stop`?  Does dccm
> switch to the second server?

Yes, `cdcc info' shows both DCC servers, and both greylist servers.
The active servers switch regularly.  If I shut down one server,
the other one becomes active.  Does that behavior indicate what
`dccm' is doing?  If so, it all seems to be working correctly.

-Gary Mills-    -Unix Support-    -U of M Academic Computing and Networking-

More information about the DCC mailing list

Contact by mail or use the form.