frequent disk IO

  • Seems like after every OMV update I need to go back and add the interval in to /etc/collectd.collectd.conf


    I found this page - http://info4admins.com/tips-to…disk-in-debian-or-ubuntu/ which led me to


    cat /proc/sys/vm/dirty_writeback_centisecs


    which by default was set to 500 (5 secs). I changed that to 90000 (15 minutes) and now my disk lights up ONCE a minute for raid sync (jbd2/md). So that got rid of the raid flush's 1-2 times per minute. Still looking to increase that.

  • I've found https://collectd.org/wiki/inde…Inside_the_RRDtool_plugin and modified the rrdtool config in OMV 0.5 to


    Code
    LoadPlugin rrdtool
    <Plugin rrdtool>
      DataDir "/var/lib/collectd/rrd"
      CacheTimeout 120
      CacheFlush 900
      WritesPerSecond 50      <========== new
    </Plugin>


    The omv-mkgraph command sends now a FLUSH command before creating the graphs, but i'm not sure if this will finally reduce disk I/O. Any tip how to improve collectd is welcome. But i will not use RRDCacheD plugin.


    VotDev,


    I upgraded to Kralizec/Wheezy and it went very well! Today I checked the system with a script:


    echo " TIME TID PRIO USER DISK READ DISK WRITE SWAPIN IO COMMAND"
    iotop -o -b -t -qqq


    and I see that the rrdcached IO is back again so I added the interval again and restarted collectd so slow it down. If this was a permanent change you made then maybe it was missed on the upgrade? Just wanted to let you know.


    Thank you and everyone who works on OMV!

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!