Trying to figure out why mysql server crashes

Discussion in 'Server Operation' started by Ovidiu, Apr 13, 2015.

  1. Ovidiu

    Ovidiu Active Member

    This is the smallest server from digitalocean, running the latest Debian distro.
    The crashes used to happen every few days until I scheduled a check via cronjob:
    0 5 * * * mysqlcheck -Aos > /dev/null 2>&1
    Now it happens quite rarely but this is what I can find in the logs and unfortunately that's all:
    Apr 12 23:11:45 server vmunix: [5590077.553606] [ 1209]   105  1209     1075       40   0       0             0 mysqld_safe
    Apr 12 23:11:45 server vmunix: [5590077.554124] [ 1663]   105  1663   247994    37235   0       0             0 mysqld
    Apr 12 23:11:45 server vmunix: [5590077.563876] Out of memory: Kill process 1663 (mysqld) score 293 or sacrifice child
    Apr 12 23:11:45 server vmunix: [5590077.564404] Killed process 1663 (mysqld) total-vm:991976kB, anon-rss:148940kB, file-rss:0kB
    I know out of memory sounds scary but what exactly is vmunix? Can't really find much about it :-(
  2. DDArt

    DDArt Member

    That is a common issue, esp. on the 5 buck droplet or even 10. Many upgrade from 5 to 10 and not tweak the mysql config file to reflect the upgrade. mysqltuner can also help. Adding swap space also might help and there's a tutorial on that on DO. You might just have to pay the 20 bucks a month to get at least 2g of ram with swap space should help quite a bit. You can also look into 1and1 but I haven't used them in a long long time and I can't speak for their tech support, last time I had them they were a UK based server.
  3. Ovidiu

    Ovidiu Active Member


    still, I can't seem to figure out what vmunix is. any clues?

    regarding ressources:
    that is obviously the easiest explanation but the server is tuned and tweaked and made to run on minimal resources. its only serving 2 very low volume websites and that's all. when I say low-volume I mean something like 10 visitors a day. it also sends out a newsletter to thousands of subscribers, all working fine. I monitor the server with munin and there are no spikes at the time it crashes...

    [--] Total buffers: 224.0M global + 1.1M per thread (50 max threads)
    [OK] Maximum possible memory usage: 277.1M (55% of installed RAM)
    The other 50% of RAM are more than enough for the other tasks running on the server.

Share This Page