Re: spamass-milter dies under heavy load

[Matt Cramer]

On Mon, 23 Sep 2002, Dan Nelson wrote:

> You're right.  I forgot SPAMD was exec()'d.  I guess a more general
> solution would be to allow the user to specify spamc args when starting
> the milter.  I wonder if there's a patch for that already.

Yes, that would be nice.  I looked on the fsf site and didn't see one in
the list of patches.  Maybe we can help in that regard but first I need
to keep spamass-milter running.  ;)

> > I am interested to see if this configuration (with spamc->spamd over
> > the network rather than sendmail->spamass-milter) improves the
> > ability to perform under load.
>
> I'd guess it would.  The milter gets called once per header, and once
> per NN bytes of body text (not sure the exact size).  Contrast that
> with spamc, which sends the entire message to spamd, and receives the
> entire response in one transaction.  Anything that reduces round-trips
> over the network will speed you up.

This is probably a better solution in the long run, but it hasn't solved
our real problem.

First of all I am not a C++ programmer.  I was a C programmer a few years
ago and can read and write a bit of C++ but not very well.  I have been
tapping one of our developers here to help me debug our problem.

Is spamass-milter thread safe?  When we run we go about 20 minutes at most
and then spamass-milter core dumps.  We are trying to go through the core
dumps and run systraces on the milter while it is running.  Is anyone else
using spamass-milt in a threaded environment like HPUX?


Matt

-- 
Matthew S. Cramer <address@hidden>          Office: 717-396-5032
Infrastructure Security Analyst                     Fax:    717-396-5590
Armstrong World Industries, Inc.                    Cell:   717-917-7099