[SOLVED] How to control process to not consume more then 20% of CPU?

[abcuser]

Hi,
using Ubuntu 8.10 I have installed VirtualBox to have multiple virtual machines. In VirtualBox tool I can set max memory consumption, but can't set max CPU consumption. I have 5 virtual computers on my physical computer and I would like to set that each virtual computer should not take more then 20% of CPU power.

Is there any way on Linux that I could control the maximum CPU consumption by each Linux process?
Regards

[aurelieng]

http://cpulimit.sourceforge.net/

:)

[abcuser]

Hi,
I have installed program:
sudo apt-get install cpulimit
and read man pages:
man cpulimit

This is nice tool. But it is not exactly what I would like to have. I need somekind of deamon that will control all of the processes and if some process (for example starting new virtual computer) try to consume more then 20% of CPU then this process is limited to use no more then 20%.

Cpulimit is nice tool for known existing processes, but I would also like to have some limitation for new processes. Any idea?

[aurelieng]

The CPU usage balance between all processes running on your machine can be adjusted by changing their "nice" value. Perhaps you should go this way ?

[abcuser]

@aurelieng, I know there is nice/renice command, but the main problem is I have to do this manually. What I would like to have is somekind of deamon that will monitor system and nice/renice (or cpulimit) the system dynamically without human intervention.

[abcuser]

Hi,
I have written bash script which behaves like a deamon. I gave it a name cpu_control.sh

EDIT: I have removed the script from this post, changed the script (repair bugs) and posted it few posts bellow.

Regards,
Absuser

[abcuser]

Hi,
I have run the above script with root privileges with nohup command, so I can safely close terminal and command will not terminate.

Run cpu_control in background:

sudo nohup ./cpu_control.sh &


Terminate cpu_control process:

ps aux | grep cpu_control ==> get pid of this command
sudo kill -9 pid_from_previous_command


Terminate all cpulimit processes (if desired):

ps aux | grep cpulimit ==> get pid of this command
sudo kill -9 pid_from_previous_command


I would be very happy if someone out there can test this script (deamon) and would be very happy if some writes some tips how to improve it.

Regards,
Abcuser

[abcuser]

Hi,
I have changed the code. The process can still occupy more then 20% of cpu if cpulimited to 20% in the interval of daemon is sleeping, so that is why I have changed the code. I have also add variables to adjust script to any other limit.


#!/bin/bash
# ================================================== ============
# CPU limit daemon - set PID's max. percentage CPU consumptions
# ================================================== ============

# Variables
CPU_LIMIT=20 # Maximum percentage CPU consumption by each PID
DAEMON_INTERVAL=3 # Daemon check interval

while sleep $DAEMON_INTERVAL
do
NEW_PIDS=$(top -b -n1 | gawk 'NR>6 && $9>CPU_LIMIT {print $1}' CPU_LIMIT=$CPU_LIMIT) # Violating PIDs
LIMITED_PIDS=$(ps -eo args | gawk '$1=="cpulimit" {print $3}') # Already limited PIDs
QUEUE_PIDS=$(diff <(echo "$NEW_PIDS") <(echo "$LIMITED_PIDS") | grep '^<' | sed 's/< //g') # PIDs in queue

for i in $QUEUE_PIDS
do
cpulimit -p $i -l $CPU_LIMIT -z & # Limit new violating processes
done
done


Now a question: how to boot up this script at start time. I have included (modified) script in /etc/init.d and executed update-rc.d command. At boot time Ubuntu freezes - it looks it hangs in while...do loop. OK, I have rescue my Ubuntu with recovery mode, but I am wondering how to write boot up script that will work without hanging. Any idea?
Regards,
Abcuser

[abcuser]

Hi,
I have managed to write boot-up script.

1. copy daemon script from above post to /root/cpulimit_daemon.sh file

2. with text editor save the following commands to file /etc/init.d/cpulimit

#!/bin/sh
#
# Script to start CPU limit daemon
#
set -e
. /lib/lsb/init-functions
case "$1" in
start)
nohup /root/cpulimit_daemon.sh &
;;
stop)
ps -eo pid,args | gawk '$3=="/root/cpulimit_daemon.sh" {print $1}' | xargs kill -9 # kill cpulimit daemon
ps -eo pid,args | gawk '$2=="cpulimit" {print $1}' | xargs kill -9 # release cpulimited process to normal priority
;;
restart)
ps -eo pid,args | gawk '$3=="/root/cpulimit_daemon.sh" {print $1}' | xargs kill -9 # kill cpulimit daemon
ps -eo pid,args | gawk '$2=="cpulimit" {print $1}' | xargs kill -9 # release cpulimited process to normal priority
sleep 5
/root/cpulimit_daemon.sh &
;;
status)
ps -eo pid,args | gawk '$3=="/root/cpulimit_daemon.sh" {print}' | wc -l | gawk '{ if ($1 == 1) print " * cpulimit daemon is running"; else print " * cpulimit daemon is not running" }'
;;
esac
exit 0


3. change file's owner to root:

cd /etc/init.d/
sudo chown root:root cpulimit

4. change permissions:
sudo chmod 755 cpulimit

5. add script to boot-up procedure directories:
sudo update-rc.d cpulimit defaults

6. reboot
sudo reboot

After reboot check:
sudo service cpulimit status
it should return: "cpulimit daemon is running" or "cpulimit daemon is not running".

Note: For users using prior to Ubuntu 8.10 instead of service command use "sudo /etc/init.d/cpulimit status" syntax or install sysvconfig package using command: sudo apt-get install sysvconfig

To stop cpulimit service:
sudo service cpulimit stop

To start cpulimit service:
sudo service cpulimit start

If you change some settings in /root/cpulimit_daemon.sh for example CPU limit to something else then 20% or daemon interval to something else then 3 seconds, then after changing settings you can restart service:
sudo service cpulimit restart

Hope this helps to someone out there.
Regards,
Abcuser

[lex1]

thankyou very much it will be very useful.

lex1:KS