Very Computer

Hello,

We are running SCO SendMail ver 8.8.8b which is working well except for
twice now we have had the system get very slow only to find the cpu idle
time at 0% and multiple popper processes running which are acruing all the
time. Once I kill them (a -9 is required) the system performance comes back
and everything works fine.

What would cause popper processes to continue running? There are multiple
users checking their mail throughout the day and all their processes don't
get stuck, there must be something causing some of them to get in a loop and
continue running.

Thanks for any information or assistance.

Later,
 John.

What version of Sco do you have and what patches are loaded?

says...

Quote:> Hello,

> We are running SCO SendMail ver 8.8.8b which is working well except for
> twice now we have had the system get very slow only to find the cpu idle
> time at 0% and multiple popper processes running which are acruing all the
> time. Once I kill them (a -9 is required) the system performance comes back
> and everything works fine.

> What would cause popper processes to continue running? There are multiple
> users checking their mail throughout the day and all their processes don't
> get stuck, there must be something causing some of them to get in a loop and
> continue running.

> Thanks for any information or assistance.

> Later,
>  John.

OpenServer 5.0.5 and Compaq's EFS 5.26a.

The system was installed with Compaq's Smart Start which installed a "UDK
Compatibility Module for OS 5.0.5 (ver 5.0.5a)", but NOT rs505a!

I'll have to get it installed which from what I've been reading is rather a
pain on Compaq machines (because of the EFS's).

Later,
 John.

Well, seeing as though the 'popper' command has nothing to do with the
Compaq EFS, and nothing to do with 'Sendmail', I beleive you're looking down
the wrong track.  A number of times on machines of various makes, the
'popper' program (which is what actually allows the user to 'POP' mail off
the server) goes rabid and chews resources, but we've never had a situation
where it actually caused any real problems (only did that for a few
seconds).  Only other time I've seen it is when the mailbox is exteremely
large (in excess of 15+Mb).  Get a few of those going, and the server just
takes a dive (this is what happens on our in-house server, mind you that's
only an OLD P90).

We've got PII 500's etc. out there with no such problems..

Might wanna track down if it's the same user causing the problem ever time,
what kind of mail they are trying to pop off the server, what email client
they are using, how big their mailbox is etc. etc. etc.

hope this gives you a bit of a head start.

Bkxie
--
Yes! I am a pixie!

Be careful when you use a kill -9 on a process as it just like
shooting it in the head and no cleanup is performed.  Only use it
if a kill -15 or kill -1 doesn't work.

A side-effect of a kill -9 and not perfroming the cleanup is that
it can leave a .<username>.pop file in place.  That is the content
of the mail box when the pop connection was made, so that new mail
can go into the now empty mailbox.  Upon exit the copy of the
dot-pop file is copied back to mbox, or appeneded.

If the kill -9 leaves the .pop in existance then the user will get
a mail-box busy/in-use/??? message when they next try to connect.
Then to make sure they don't miss any mail, and that you don't
clobber incoming mail, you need to turn of the MTA, append the
dot-pop file to the current mbox content and then restrat the MTA.

kill -9  is the worst of all possible deaths.

--