Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Constant load ( ~1.00) on server
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Constant load ( ~1.00) on server

wroxwrox Member
edited February 2013 in Help

Lately, I've been noticing some strange CPU load averages on my unused Kimsufi box. Having nginx and php5-fpm only installed, it varies from 1.00 to 1.08, is consistent, and has lasted for some days now even with the both mentioned services stopped.

Does anyone have any idea on how to find the source of the load and remove this burden? Thank you in advance!

-EDIT: My apologies! This thread should be in Help.-
Changed the category. Thanks for the tip!

Comments

  • jhjh Member

    top

  • IshaqIshaq Member
    edited February 2013

    You can change the category, it's fine.

    Regarding the server, try installing htop (apt-get install htop, or yum install htop) and using the command "htop" this is a good tool to find out what kind of processes are using the CPU.

    Is this a dedicated box? it could be old hardware.

  • wroxwrox Member
    edited February 2013

    @jhadley @Ishaq I have tried that and htop, which is what I regulary use.

    Neither shows the source - 0.0 load on most if not all instances. Strange!

    Yes. It's the KS R-8G with an Intel Core i3 2130 CPU. Not sure if it's to be viewed as old.

    Thank you for the replies!

  • Harddrive failures?

  • IshaqIshaq Member
    edited February 2013

    Do you have a lot of memory used or something constantly updating? database perhaps? I/O causes load.

  • Run top, see which process is stuck?
    You either have a process stuck in "R" state (Running) or "D" state (waiting for an i/o operation to complete).

  • jhjh Member

    @wrox said: Neither shows the source - 0.0 load on most if not all instances. Strange!

    What does it show when you press 1?

  • wroxwrox Member
    edited February 2013

    @taronyu I appreciate that perspective being taken up for the sake of solving the problem, would you mind elaborating on the issue?

    @Ishaq About 225mb right now. I have MySQL installed, but it is stopped. No crons either. Makes me wonder..

    @jhadley http://i.imgur.com/ymvM1Sr.png

    EDIT: htop: http://i.imgur.com/oWqcWzb.png

  • 75% wa - you have something stuck in "D" state.
    Do ps axuww | grep " D "

  • What's that 225mb doing?

  • Are you running the 3.2.13-grsec-xxxx-grs-ipv6-64 kernel? (use uname -a to find it out)

  • Ok, the htop says you have a stuck "sftp-server".
    Now you have to figure out why it's stuck and whether you can kill it.

  • wroxwrox Member
    edited February 2013

    @rds100 Sure! I'm not sure what this means.
    root@reducted:~# ps axuww | grep " D "
    root 1353 0.0 0.0 10140 840 pts/0 S+ 20:19 0:00 grep D

    @Ishaq I presumed it was nginx eating up all that memory, but now when I think about it.. I'm not sure.

    @gsrdgrdghd I am running the 3.2.13-xxxx-std-ipv6-64-hz1000 kernel.

    @rds100 It might be because of a couple 100gb-transfers to the server not so long ago, via sFTP, which put excessive load on the server (possibly making it bugging up, no idea). Giving the sftp-server a SIGKILL doesn't seem to work. Any ideas?

    Thank you everyone!

  • @wrox said: Giving the sftp-server a SIGKILL doesn't seem to work. Any ideas?

    ls -l /proc/11035/fd
    to see what it's trying to read/write.

  • wroxwrox Member
    edited February 2013

    @rds100
    lr-x------ 1 root root 64 Feb 4 13:22 0 -> pipe:[300956]
    l-wx------ 1 root root 64 Feb 4 13:22 1 -> pipe:[300957]
    l-wx------ 1 root root 64 Feb 4 13:22 2 -> pipe:[300958]

  • You can't kill a process in D state. It's waiting for I/O. Usually it's a bad RAID/drive or NFS mount gone bad. What's in dmesg?

  • @wrok now pstree -p (and see which process started the sftp-server and which processes were started by it).
    Killing the childs / parent might help.

  • wroxwrox Member
    edited February 2013

    @t3k9 I see. Meaning, it is locked until it gets what it wants?
    http://i.imgur.com/5DICN0R.png That all over it seems..

    @rds100
    It doesn't seem like it has any parent nor child processes. Here you go! http://i.imgur.com/hxKkHv5.png

    This is certainly interesting.

  • Is it an option to just reboot the server?

  • @gsrdgrdghd said: Is it an option to just reboot the server?

    Same question. That would most likely fix your issue.

  • Third on the reboot but might want to check the software RAID.

  • Rebooted and it's all back to normal now. Thank you guys!

    @t3k9 If it occours again, I shall. I currently don't have the knowledge to do that especially since OVH set that up for me automatically.

  • @wrox said: Rebooted and it's all back to normal now. Thank you guys!

    @t3k9 If it occours again, I shall. I currently don't have the knowledge to do that especially since OVH set that up for me automatically.

    To start with run:

    cat /proc/mdadm

    It will show you the array status.

Sign In or Register to comment.