Work at SourceForge, help us to make it a better place! We have an immediate need for a Support Technician in our San Francisco or Denver office.

Close

Disk monitoring misread

Help
Steffen
2013-07-12
2013-07-14
  • Steffen
    Steffen
    2013-07-12

    Hi guys.

    Recently I experienced that xymon began to show error on various hosts' disk monitoring. It seems that the report sent by the client has been misread in some way which will give an incorrect information and let the status of the client go red.

    I've tried to past in the report below and hightlighted some of the lines which looks wrong to me.

    6% / (19917908% used) has reached the PANIC level (95%)
    1% /dev (500400% used) has reached the PANIC level (95%)
    1% /run (203596% used) has reached the PANIC level (95%)
    0% /run/lock (5120% used) has reached the PANIC level (95%)
    0% /run/shm (509744% used) has reached the PANIC level (95%)
    2% (2060992% used) has reached the PANIC level (95%)
    11% /boot (940456% used) has reached the PANIC level (95%)
    2% /var (52509252% used) has reached the PANIC level (95%)
    1% /var/log (10417164% used) has reached the PANIC level (95%)

    Filesystem
    102]
    blo]ks
    Used Available Capacity Mounted on
    /dev/md4 20969408 1051500 19917908 6% /
    udev 500404 4 500400 1% /dev
    tmpfs 203900 304 203596 1% /run
    none 5120 0 5120 0% /run/lock
    none 509744 0 509744 0% /run/shm
    /dev/md2 2094080 33088 2060992 2%
    /tmp

    /]ev/md0 1049600 109144 940456 11% /boot
    /dev/md5 53405184 895932 52509252 2% /var
    /dev/md3 10479616 62452 10417164 1% /var/log

    After a short while another report will be recieved and read correctly and the status goes green. Usually it has never been a problem and as told it happens for different clients on two hobbit servers. The fun thing is, that these two servers report disk errors different from each other. If one server reports an disk error for one client, the other server might not show it and the other way around.

    Hope you can point me in any direction to solve this.

    /Steffen

     
  • Steffen
    Steffen
    2013-07-14

    To your information I began to notice some messages regarding too big client reports in the hobbit logs - not the same clients as mentioned above, but one single client. After adjusting the variables:

    /etc/hobbit/hobbitserver.cfg

    MAXMSG_STATUS=""
    MAXMSG_DATA=""
    MAXMSG_CLIENT=""

    To suiting values, the reports seems the be read fine again.