Unusual # of MailScanner processes running

[Jason Burzenski]

Hi all,

I am monitoring an unusual situation on one of two identical MS boxes.  Both
boxes are RH9, running MS 4.29.7 (upgraded on 4/3/04).  These boxes have
been running with symetric configurations for over 200 days without as much
as a hiccup.

Starting at 12:30PM yesterday (4/15), one of the two boxes started running a
higher than usual number of MS processes.  I noticed the issue from
mailscanner-mrtg which has shown almost a flatline for # of MS copies
running since the box was turned up (normally run 10 copies).  Starting at
12:30PM, the graph bounces between 10 and 18 copies continuously (producing
a much ulglier graph).  I have confirmed these are actual copies and not a
script with a similar name that is being picked up by msmrtg.

I've reviewed time slice of logs (several thousand lines) from when the
problem occurred and logs during the unstable operation and have found not
even a clue as to what might be causing these extra processes.  The
identical box still shows a flat graph.

The load on these boxes is nearly identical as well (~45Kmsgs/day, 50%spam,
~2K virus).  I am not noticing any performance issues on either box (other
than slightly increased #ofmsgs in the queue.in on the problem box), even
with the bouncing process count but it is bugging me that I cant figure out
whats wrong and why this would happen independent of any changes I've made.


The changes I've made since the upgrade were the put a ruleset on the allow
scripts tag to permit 3 domains.  I've confirmed the MailScanner.conf and
the rules file I used are exactly the same on both boxes.

Has anyone experienced anything like this in the past?  Anyone have any
ideas?  I may go debug on the system this weekend but I'll have to schedule
the downtime.

Thanks,

Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mailscanner.info/pipermail/mailscanner/attachments/20040416/0f7e2086/attachment.html