*BSD News Article 43404


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!uwm.edu!spool.mu.edu!howland.reston.ans.net!news.sprintlink.net!bga.com!bga.com!nobody
From: dpm@bga.com (David P. Maynard)
Newsgroups: comp.unix.bsd.bsdi.misc
Subject: Cron "No more prcesses" spawning nntpsend
Date: 14 Mar 1995 10:56:36 -0600
Organization: Real/Time Communications - Bob Gustwick and Associates
Lines: 34
Message-ID: <3k4hs4$h9q@maria.bga.com>
NNTP-Posting-Host: maria.bga.com


The cron process on our news server occassionally emits the following
mail messages:

Subject: Cron <news@news> /usr/local/news/bin/nntpsend >& /dev/null
No more processes.

The messages are clustered in time.  I noticed that the last bunch
appeared not long after one of our busiest systems crashed and
rebooted, so it is probable that a lot of processes were "left
hanging" waiting on a timeout while a new batch of nnrpd processes was
created.

The question is,.... which process limit are we hitting?  The message
is generated by cron, so I don't think that innd is directly involved.
(I already increased the user process limit to 128 in the rc.news
file.)  I have maxusers set to 64 which yields an NPROC value of 1104
and I know there aren't 1100 processes.  

I checked all of the crontab files and I doubt that cron has 64
concurrent processes at any time.  If it does, there is something
"wrong."  I've never seen any sign that there were a large number of
cron jobs hanging around though.

Any suggestions?

Thanks.

-dpm


-- 
 David P. Maynard, dpm@depend.com
--