High cpu load during the night..

[Duboux]

I noticed a high serverload (80%) during the night..

I bet it is some cron job..
But I can't find a way to see all the cronjobs that are on that server..

What is a bit disturbing is that this high load takes over an hour.

The box has a dual core 2x2.80GHz and 1gb of ram.

How can I see allll the cronjobs in 1 view ?

 

[nobaloney]

Do you have a scheduled backup? That's the most obvious possibility.

Jeff

 

[Duboux]

I don't know how to see all the cronjobs, but when I type:

# crontab -l
# backup
0 1 * * * /*****/backup.scp
# Copy yesterdays exim reject logs to a database
0 2 * * * /*****/cron_eximreject_2_db.sh

There's a backup proces alright, but it's at 1 am, not at 5.

How can I find other cronjobs ?

# ls -l /etc/cron.daily
total 68
lrwxrwxrwx  1 root root   28 Mar  3 17:05 00-logwatch -> ../log.d/scripts/logwatch.pl
-rwxr-xr-x  1 root root  418 Nov 19  2004 00-makewhatis.cron
-rwxr-xr-x  1 root root  276 Sep 28  2004 0anacron
-rwxr-xr-x  1 root root 2133 Nov 23  2004 prelink
-rwxr-xr-x  1 root root  104 Nov  2  2004 rpm
-rwxr-xr-x  1 root root   82 Aug 10  2005 slocate.cron
-rwxr-xr-x  1 root root   58 May 20 18:00 tijd [COLOR="Red"]<-- I made this one, to update the time and date[/COLOR]
-rwxr-xr-x  1 root root  286 Aug 14  2004 tmpwatch
-rwxr-xr-x  1 root root  136 Jul  5  2005 yum.cron

 

[nobaloney]

I don't know what your backup.scp job is. I don't have one. I have a copy of sysbk running in my root user cronjob. sysbk can add a lot of serverload; perhaps your backup.scp job does, too.

What time does your system run the jobs in /etc/cron.daily (you can find that out by looking into /etc/crontab).

$ crontab -l

is a look into the current user's crontab. You can run it for other users this way, but only as root. For example:

# crontab -l -u admin

for the admin user, and for other users as well. But that won't tell you which users have crontabs.

The list of users with crontabs is in /var/spool/cron. Look for non-zero-length files. You can just use an editor (or the less command) to look through the cronjobs in /var/spool/cron. But don't make changes to these files manually; instead edit them with the crontab command:

# crontab -e -u admin

The above will let you edit the crontab for admin using the VIM editor, or alternatively whatever editor is set in your EDITOR environment variable.

Have I managed to be even more confusing ?

Jeff

 

[Duboux]

Thanx for the help

The crontab -l (or -e) command showed the 2 cronjobs I posted above (the backup and the eximthing)
there's no cronjob on admin, since there's no admin account on that server. (well there is, but no domain/useraccount)

# cat /etc/crontab
SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin
MAILTO=root
HOME=/

# run-parts
01 * * * * root run-parts /etc/cron.hourly
02 4 * * * root run-parts /etc/cron.daily
22 4 * * 0 root run-parts /etc/cron.weekly
42 4 1 * * root run-parts /etc/cron.monthly

This 4am daily cron probably explains the little spike in the cpu-load-image, posted earlier

# ls /var/spool/cron
diradmin  user1 user2  root  user3

I found a cronjob that starts around 5am..

# crontab -l -u diradmin
5 5 * * *  echo -e "action=backup&id=1&owner=admin" >> /usr/local/directadmin/data/task.queue.da; /bin/chmod 600 /usr/local/directadmin/data/task.queue.da

So this is a backup ?

 

[nobaloney]

It looks like a backup.

Since it appears to be an admin backup you should look at all backup possibilities in the admin level of the control panel.

Jeff

 

[hostpc.com]

If you have root access, create this script as readcrontab.sh

#!/bin/sh

myemail=$1
if [ -n "$myemail" ] ; then
        mycrontmp=/root/cron.tmp.$$
        # get current user list
        for i in `cat /etc/passwd | cut -f1 -d ':' | grep -v '#'`; do
                echo "--------------------------------------------------"
                echo "Username: ${i}"
                echo "--------------------------------------------------"
                crontab -u ${i} -l 2>&1
                echo "--------------------------------------------------"
        done > $mycrontmp
        cat $mycrontmp | mail -s "crontab report for `hostname`" ${myemail}
        rm -f $mycrontmp
else
        echo "Please supply a valid email address to get the report."
        exit
fi

chmod 755 readcrontab.sh

then ./readcrontab.sh [email protected]

it'll send an email of ALL user crons on the server

 

[nobaloney]

Nice script, Joe.

Coincidentally we had the same problem with a system last night; it appears to be caused by the sysbk backing up a huge site. I don't know why that would do it, unless the nice level set in conf.sysbk isn't working, but that's what it looked like to us.

Jeff