[cobalt-users] Sendmail

["Ryan Howe" <ryan@xxxxxxxxxx>]

Hello,
	I was wondering if anyone could give me some tips or point me to some sites
that would have tips on blocking spam.  I don't want a spam filter per say
(we have a third party handle the filtering).  Let me explain the problem.
Right now our Raq 4 has about 500+ emails in the mail queue that are
deferred spam to bad email addresses (at least that is what it looks like),
also we have about 225 sendmail processes running.  The number of email in
the queue has not dropped below 400 in the last three weeks.  This server is
running dog slow, it takes it over 1 min or more just to verify  a password
when ssh'in to it.   The server is really being over worked, top's output is
as follows.

3:04pm  up 1 day, 18:43,  3 users,  load average: 5.16, 5.20, 5.85
317 processes: 313 sleeping, 4 running, 0 zombie, 0 stopped
CPU states: 44.6% user, 50.9% system,  4.4% nice,  0.0% idle
Mem:   517104K av,  513620K used,    3484K free,  516952K shrd,  127196K
buff
Swap:  131532K av,      12K used,  131520K free                   86856K
cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME COMMAND
 3324 root      18   0   544  544   444 R       0 29.0  0.1  51:38 syslogd
22570 gpeople   10   0  5912 5912  1260 R       0 19.7  1.1   0:02
.openwebmail-ma
22569 root      10   0  1048 1048   676 R       0  9.8  0.2   0:02 top
22902 root      12   5 13004  12M  1884 S N     0  3.8  2.5   3:37
procqueue.pl
22913 postgres   7   5  2980 2980  2296 S N     0  0.8  0.5   1:05
postmaster
18890 root       0   0  1844 1844  1244 S       0  0.5  0.3   0:00 sendmail
27402 root       1   0  2376 2376  1240 S       0  0.4  0.4   0:04 sendmail
 2176 root       0   0  2080 2080  1264 S       0  0.4  0.4   0:02 sendmail

Mail is becoming a real problem on this server, it becomes so sluggish that
it needs to be rebooted every few days.  Also it is filling the /var partion
because the maillog file is getting so large that even log rotate fails so
we have to monitor it very closely.  Also syslogd seems to be eating
anywhere from 29-80% of the cpu at any time (usually averaging around 50%).
I think that sendmail is waiting for a connection with "lame" servers for
like 2 minutes.  Is it advisable to change this to somewhere to around 1 min
or is that a bad idea? Most of the emails that are sitting in the queue are
from MAILER-DAEMON. Ex:

3223 Mon Oct 20 15:13 MAILER-DAEMON
(Deferred: Connection refused by offersjust4you.com.)
<ev06tmapqdb@xxxxxxxxxxxxxxxxxx>

I have already changed sendmail to remove all mail that is undeliverable for
2 days, I don't want to go any lower than that.  I want to work through
these problems issue by issue and right now I want to solve the mail issue.
So any advise would be greatly appreciated.

Thank you

Ryan