[cobalt-users] Re: Very Slow RAQ4

[Peter Frederick <pfred@xxxxxxxxx>]

On Wed, 29 Jan 2003 15:23:07 -0000, you wrote:

>Everyone seems to have run out of ideas - little
>bit more information, we've started receiving a
>load of old emails (2 weeks old) that we've already
>had. Could the two situations be connected?
>
>

See how many sendmail daemons are running by typing

ps -aefl | grep sendmail | grep -v grep

You may have a bunch of them! 

We had this issue on a Qube3 - we had a site connecting to us via SMTP that
has a protocol issue and the sendmail daemon that was being fired up would
not finish (It was expecting data from the other end). This would create
two processes liked (via PPID) to the original sendmail process (It's
parent ID is 1).

Since the mail delivery didn't work, the bad site would try again (before
the timeout occurred). At one point I counted 52 sendmail processes in
existence - all related to that same site.

In the end I added that IP address to our access database to block them
until the problem can be resolved.

The machine was running very slowly if it builds up that many sendmail
processes. I believe you can adust some timeout values and max number of
daemons in the sendmail.cf but I haven't got that far.

Maybe you have a problem like this.

What does your machine load show?

ie:   w

Cheers
Peter


Peter Frederick
MIS Director, Indiana Packers Corp, Delphi IN
Phone: (765) 564-9705   Fax: (765) 564-3684
Work: pfred@xxxxxxxxx   (Qube3 Professional running 6.4)
============================================================================
'Tis better to light one candle than to curse the darkness a thousand times!