Re: [cobalt-users] swatch taking 90% of cpu

["Dave" <maxdoubt@xxxxxx>]

----- Original Message -----
From: "Chris Heiner"
Subject: [cobalt-users] swatch taking 90% of cpu


> How can I remove this service completely? I have "killed" this service many
> times a day. And it keep s coming back.

I had this problem on a Raq3.  As it turned out for me, the postgres was
damaged.  One way to tell if this is the case for you as well is to ssh into
the machine and run TOP.  Look for CROND all in caps - this is a second
instance of cron running.  If this IS the case for you then you have two
options:  1) Call Cobalt to rebuild postgres or 2) TRY THIS at your OWN RISK
as it has worked for a few people:

su to root
kill swatch and CROND (the CAPS one)

edit /etc/crontab by inserting a # in front of
0,15,30,45 ****root /usr/sbin/swatch>>/var/cobalt/adm.log 2>&1  (Raq 3,4 and
XTR)
so Active Monitor won't restart stopped services...

Kill the sendmail process:
/etc/rc.d/init.d/sendmail stop

cd /home/pgsql
su postgres
edit pg_hba.conf
change crypt to trust
pg_dump cobalt > cobalt.dump
vi (or pico) cobalt.dump
find create_table for quota and hash it out ( comment it out )
close
destroydb cobalt
createdb cobalt
psql cobalt < cobalt.dump
edit pg_hba.conf
change trust back to crypt

edit /etc/crontab and remove the # in front of
0,15,30,45 ****root /usr/sbin/swatch>>/var/cobalt/adm.log 2>&1

Restart sendmail:
/usr/local/sbin/swatch

You're all done....please excuse me for posting the full solution if this IS
NOT your problem.  I just wanted to make sure you had the instructions in
case....

Max~