I'm sure this is contributing to the problem as most of us are under
dictionary attacks at some point in the day. My concern is that the smtp
service isn't releasing the memory.
Dan Horne wrote:
Could this be the result of excessive dictionary attacks? Even with
8.2's new feature to protect from these, if the attacks are distributed
across many IP addresses it could still bog down your server. When it
was happening to us, I didn't check smtpd32's mem usage, but I do know
that our server needed a reboot almost daily because it would slow down
until it couldn't process any more mail at all. Search your logs for
"invalid user" and see how many hits you get. Just a thought.
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Matrosity Tech
Support
Sent: Wednesday, May 18, 2005 4:55 PM
To: [email protected]
Subject: Re: [IMail Forum] SMTPD Memory usage
No! good
Which version are you running? 8.2 HF1
What OS? Win2K SP4
What kind of hardware? Dell
How long does it run before needing a reboot? not consistent but ram
usage grows above 100 MB's in about 2 hrs
for 8.2 you should see memory between 6 and 20 mb for smaller installs
but it can grow to 60 - 120 mb for a very very active server where many
users are sending very large attachments. But the mem usage should drop
back down to between 10 and 20 mb during low usage times.
memory never drops down
BTW, is there a 'nobody' alias on one of your domains? Or do you have a
lot of mail going to one mailbox? This could cause temporary spikes in
mem usage if that account is receiving mail with a lot of attachments.
only on our domain
Another issue that I have seen behavior like that in testing 8.2 is the
spool drive needing a chkdisk to clean up bad indexes which can be
caused by killing a process while it is creating or deleting a file.
The issue is that if the service cannot create the spool files the
threads will block trying to do that and you will see memory grow, the
thread count grow beyond your Max setting (default 60) and the server
once all the worker threads are blocked, does not do anything until you
kill it. But killing it does not solve the problem, running chkdsk /r
does in this case. Defragging is good and should be done often but does
not address this particular issue.
We're degragging every 4 hours
To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive:
http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html
List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/
Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
