[Mimedefang] Scaling MIMEDefang / SA on Solaris

[Joseph Brennan]

> I'm sure that I'm facing resource problems.  I've read that Solaris 8
> 32-bit applications are faced with a file descriptor limit of 1024 with
> milter's use of select(), regardless of how high this tunable is set in
> /etc/system. I've tried limiting SMTP daemon connections, I've tried
> small, medium, and large values for MD resource pool.  I've tried
> increasing MD's BUSY timeout and I've tried up'ing the timeout values in
> the milter m4 entry.

This sounds familiar.  Try to  ls -R /var/spool/MIMEDefang/ and
see how many files are there.  Is it even 100?  And remember each
perl process gets 256 or 1024 file handles.  Forget the load, but
run top and look at the iowait.

We found that the problem was slowness reading and writing local
disk.  How slow this is depends on hardware and even Sun software.
But you probably can't get it fast enough.  We just tested making
/var/spool/MIMEDefang/ be TMPFS (ram disk) and the i/o problems
went away.  This is not in from production just yet so I can't claim
experience beyond a few hours, but others here said it was the
thing to do.


> On larger sites, I find it key to run sendmail in daemon mode only (ie
> queue only) then use persistent queue runners to process mail.  Can I
> configure MD to only run from the queue runners and not from sendmail -bd
> instances to limit MD resources?

This is slightly off-topic, but I tried persistent queue runners
and performance was was terrible.  The load was down all right,
but the persistent queue runners had to complete the whole queue
before they would start the next round, and it can take well over
an hour to run the queue with accumulated messages to unresponsive
external hosts.  Maybe I missed something.


Joseph Brennan          Columbia University in the City of New York
postmaster at columbia.edu                 Academic Technologies Group