[Eug-lug] Server load problem revisited - my solution

[Matthew Jarvis]

A while back I asked for help trying to determine why I would have a 
server that would, over time, just start going nuts with CPU load and 
requiring a reboot. For a while now I've been rebooting this thing as a 
routine precaution pretty much weekly.

I never have figured out what the overall issue is, but me thinks I've 
come up with a way around it.

My theory is that the thin client machines in our production line are 
the culprit, and the machines just being shut down rather than a 
graceful logout might be leaving dangling processes and such.

So I decided to try and just nuke everybody in a cron job every night. 
What the real issue may be I don't know (nor have the time to figure 
out), but maybe my workaround will be to just do some housecleaning instead.

So, I'm running this command in cron for each of the workstations:

15 22 * * * kill -9 `ps aux | grep workstation | awk '{print $2}'`


Thought I'd share this in case this obvious solution could benefit 
someone else.


-- 
Matthew S. Jarvis
IT Manager
Bike Friday - "Performance that Packs."
www.bikefriday.com
541/687-0487 x140
mattj at bikefriday.com