Re: [cobalt-developers] Replacing sendmail

[Jeff Lasman <jblists@xxxxxxxxxxxxx>]

"Robert G. Fisher" wrote:

> On Sun, Jul 23, 2000 at 07:49:19PM -0700, Jeff Lasman wrote:
> > One of my clients is overloading sendmail big time with large (fully
> > opt-in) mailings (they're a Hollywood movie studio).
> 
> How are they maxing it out?

I'll append a copy of the post to mailhelp@xxxxxxxxxxxxx to the end of
this post.

> Sending lots of individual messages?
> a lot of the same messages?

They're running SubscribeMe, which is a (non-open-source) mailing list
package.



> Would it be more helpful if they
> were using a mailing list?

They are.  But it shouldn't really matter.  All mail list programs I
know of (including majordomo and subscribeme) use sendmail.

-------- Original Message --------
Subject: Overloaded server?
Date: Sat, 22 Jul 2000 11:22:20 -0700
From: Jeff Lasman <jblists@xxxxxxxxxxxxx>
Reply-To: mailhelp@xxxxxxxxxxxxx
Organization: nobaloney.net
To: mailhelp@xxxxxxxxxxxxx

Can anyone help with this one...

I've got a client who's got a Cobalt RaQ3.  Last night he started
running "SubscribeMe" to send an email to a list of 17,000 messages.

This morning he got a message from the RaQ3s monitor telling him that
sendmail wasn't accepting messages.

It's not.

Here's a partial of what "ps aux | grep sendmail" shows:

root      2576  0.1  0.4  2088 1144 ?        S    11:08   0:00 sendmail:
rejecti

If I restart sendmail (# kill -HUP 2576) I get

root      2577  0.1  0.4  2088 1144 ?        S    11:08   0:00 sendmail:
accepti

for a while, but later it becomes

root      2577  0.1  0.4  2088 1144 ?        S    11:08   0:00 sendmail:
rejecti

again.

Nothing in the bat book to indicate what's going on.

Here's a top:

 11:06am  up 37 min,  1 user,  load average: 0.21, 0.28, 0.26
44 processes: 43 sleeping, 1 running, 0 zombie, 0 stopped
CPU states:  0.7% user,  1.9% system,  0.0% nice, 97.2% idle
Mem:  257964K av,  81192K used, 176772K free,  79976K shrd,  15420K buff
Swap: 131536K av,      0K used, 131536K free                 35764K
cached


Here's a df:

# df
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/hda1               743466    489064    254402  66% /
/dev/hda3               198601     17717    180884   9% /var
/dev/hda4             13380124    872507  12507617   7% /home

The mqueue directory has 2683 files in it.

We've tried rebooting the server as well, since email the RaQ3 sent us
recommended that.  Didn't work (of course <wry grin>).

We think this is simply an overload condition.

Any other ideas?

Jeff
-- 
Jeff Lasman <jblists@xxxxxxxxxxxxx>
nobaloney.net
P. O. Box 52672
Riverside, CA  92517
voice: (909) 787-8589  *  fax: (909) 782-0205