On Sun, May 12, 2002 at 12:24:40PM -0500, Richie Laager wrote:
> On Sunday 12 May 2002 11:57 am, Michael Stenner wrote:
> >  The only times I KNOW it died were at sunday morning log
> > rotation.  Granted, some other things happen then, but I
> > think the log rotation is the best bet.
> 
> I would recommend that you start experimenting. For example, 
> try restarting the syslog daemon with; /etc/init.d/syslog-ng 
> restart >>/dev/null

All good suggestions (snipped).  A little background, then my
findings.

This is a semi-serious machine.  It's the mail/nfs/nis/dns/log server
for the Duke Univ. Linux Users Group.  Several (maybe 25) do their
primary mail through it, so I don't want to muck with it a lot.  Only
1 real person is using spamd on it though.  It's being tested there to
see if we want to use it in Duke Physics.

So, I suspected that it was not dying because logs were rotated
_WHILE_ a message was being processed.  (There's just not that
much... odds were very small it would happen two weeks in a row.)

Here's what I did:
1) verify spamd is running (yes)
2) restart syslog-ng
3) verify spamd is running (yes)
4) send a test message through spamc
5) verify spamd is running (NO!!!)

The message arrived, slthough it was not processed (no SA headers).
There was nothing in the logs from spamd or spamc (presumably because
spamc connected successfully, but but spamd died before finishing).

> No matter what you find out, make sure to file a bugzilla 
> report with as much detail as possible.

I will most certainly do that.  At this point, I think we're close
enough that I think I'll just try and nail it down so I can send you a
patch along with the bug report.  If you have any further insight as
to where I should start looking, though, it would be much
appreciated.

I think I'll probably try and grab the (maybe 50 lines of) code that
does the logging, and see if I can recreate the problem on my home
server, which I'm considerably more willing to futz with.

                                        -Michael

-- 
  Michael Stenner                       Office Phone: 919-660-2513
  Duke University, Dept. of Physics       [EMAIL PROTECTED]
  Box 90305, Durham N.C. 27708-0305

_______________________________________________________________

Have big pipes? SourceForge.net is looking for download mirrors. We supply
the hardware. You get the recognition. Email Us: [EMAIL PROTECTED]
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk

Reply via email to