[Gllug] What happened to make my server run all those processes?

[John Winters]

When I went to check my e-mail this morning I found my server was
exceedingly sluggish.  A quick check indicated why - it had about 1400
processes running (compared to a normal figure of 100) and a load
average of about 800 (compared to a normal figure of about 0.2).

I'm still trying to figure out what had happened.  Unfortunately I
wasn't able to get a list of the running processes before I re-booted
it, but there didn't seem to be anything odd there - just vast numbers
of the same processes.

The graphics generated by munin indicate that things started to stray
from normality at about 22:00 yesterday.  Both the number of processes
and the load average increased in a beautifully straight line way for 8
hours through the night until I checked it at 06:00.  The logs don't
show up any odd error messages, except for the fact that my secvpn
connection (I'm the passive end) seems to have bounced every minute or
so through the night.  When I came to the system this morning, the user
"secvpn" was logged in many dozens of times (didn't manage to count
them) instead of the usual once.

I've checked the log files carefully but can't see anything which might
have triggered the events.

Anyone ever experienced anything like this before?  Is there a process
failure which would cause other processes to keep replicating?

TIA,
John

-- 
John Winters, Wallingford, Oxon, England
i = (free (NULL); i++);

-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug