Re: Excessive sendmail processor usage following 3.0 patch

["Richard Webb" <rcsw_(_at_)_periscon_(_dot_)_com>]

I think the sendmail SSL crypto stuff is a bit too much for my little
sparc, I've edited the makefile to disable this. Compiling now, so I'll
see if that helps.

R.

-----Original Message-----
From: owner-misc_(_at_)_openbsd_(_dot_)_org [mailto:owner-misc_(_at_)_openbsd_(_dot_)_org] On Behalf
Of OpenBSDList
Sent: 05 March 2003 17:52
To: 'Richard Webb'; misc_(_at_)_openbsd_(_dot_)_org
Subject: Re: Excessive sendmail processor usage following 3.0 patch


I'm experiencing a similar scenario... 3.0, i386, using the same patch.

The box is being used to relay internet mail to our internal Exchange
server.  All mail is destined for the same Exchange machine (which seems
to be working fine).  99% of mail comes in no problem, but every hour or
so a message times out with 

(Deferred: Connection timed out with [exchangeIP])

or

(timeout writing message to [exchangeIP]: Connection timed out)

This while hundreds of other messages are being delivered to ExchangeIP
with no problems.  When sendmail periodically processes the queue, the
newly spawned sendmail process(es) take up extreme amounts  of CPU
before finally being killed manually.  The messsage delivery fails with
the same error.

Any help or hints appreciated.

Mike

-----Original Message-----
From: Richard Webb [mailto:rcsw_(_at_)_periscon_(_dot_)_com]
Sent: Wednesday, March 05, 2003 12:31 PM
To: misc_(_at_)_openbsd_(_dot_)_org
Subject: Excessive sendmail processor usage following 3.0 patch


After patching vanilla 3.0 sendmail sources on obsd 3.0 (sparc LX) 
with Tor Houghton's patch for the recent ISS advisory - the new sendmail
binary Seems to hog all my cpu usage, although I can still send/receive
mail through it.

Pointers to what might be going on, and how to rectify it, would be
useful?

Rich.