What to look for to trace cause of OOM killer on a new server

Discussion in 'Server Operation' started by Ten4, Oct 3, 2013.

  1. Ten4

    Ten4 New Member

    Hi all,

    I'm relatively new to this server business. I have set up a few without any hiccups but now I have an OOM killer problem on two servers. Both of them are on Digitalocean and both of them are set up according to this tutorial: http://www.howtoforge.com/perfect-server-debian-wheezy-apache2-bind-dovecot-ispconfig-3

    The only difference is that I've replaced Squirrelmail with Roundcube and that didn't cause the problem since I installed Roundcube on the second server after I had encountered the OOM problem.

    The thing is we have another server set up by someone else from which we are migrating out of. The old machine is Ubuntu 12.04, 512 RAM, 40GB disk and the new ones are both Debian Wheezy, 512, 20GB. We had six low-traffic sites on the old server and on the new ones the other has two of those sites and the other has four. On the old server we have Webmin/Virtualmin and of course ISPConfig 3 on the new ones. The rest of the setup is identical.

    On the old server we didn't have any memory related problems. Now OOM killer kicks in on both servers about 1-2 times a day.

    I have ssh access to the old server, but not to the person who set up that box. The question is, where to look for differences in the configurations?

  2. ioFlare

    ioFlare New Member

Share This Page