pattern for check hard disk failure - Hardware

This is a discussion on pattern for check hard disk failure - Hardware ; Hi, We are developing a software which can report hard disk failure. We plan to periodicly check dmesg output and log files under /var/log to report hard disk failure. The pattern we plan to use is "hard disk failure", "hard ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: pattern for check hard disk failure

  1. pattern for check hard disk failure

    Hi,

    We are developing a software which can report hard disk failure. We
    plan to periodicly check dmesg output and log files under /var/log to
    report hard disk failure. The pattern we plan to use is "hard disk
    failure", "hard disk error". And we have one RAID0 disk for large data
    storage. For RAID drive, use pattern "RAID drive failure", "RAID drive
    error". Is those correct? Any other patterns?

    Thanks,
    Roger

  2. Re: pattern for check hard disk failure



    On Tue, 3 Jun 2008, Roger wrote:

    > Hi,
    >
    > We are developing a software which can report hard disk failure. We
    > plan to periodicly check dmesg output and log files under /var/log to
    > report hard disk failure. The pattern we plan to use is "hard disk
    > failure", "hard disk error". And we have one RAID0 disk for large data
    > storage. For RAID drive, use pattern "RAID drive failure", "RAID drive
    > error". Is those correct? Any other patterns?


    You should take a look at smartd.

    If you want to look for problems with Linux software RAID sets, then you
    should look at /proc/mdstat


  3. Re: pattern for check hard disk failure

    Roger wrote:
    > Hi,
    >
    > We are developing a software which can report hard disk failure. We
    > plan to periodicly check dmesg output and log files under /var/log to
    > report hard disk failure. The pattern we plan to use is "hard disk
    > failure", "hard disk error". And we have one RAID0 disk for large
    > data storage. For RAID drive, use pattern "RAID drive failure", "RAID
    > drive error". Is those correct? Any other patterns?


    Roger,

    smartd the too I suggest for automating this.

    I would suggest using:
    /dev/hda -H -s (S/../../1/12|L/../01/2/12) -n standby,q -m root

    As a starting point for periodic checking.

    The -m part will email your root account, although there is no reason
    why you couldn't email any internet or intranet email address, depending
    on your configuration and needs.

    Please remember though that the smart standard (and tests) are not 'good
    enough' to let you know in all circumstances that a disk is likely to
    fail soon.

    On the other hand, if however you do receive an email indicating that
    the drive thinks it will fail soon, you can put your faith in it that it
    will

    --
    Regards,
    Sheridan Hutchinson
    Sheridan@Shezza.org


    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFIRblInBrliHqz8aARAvINAKDBISDUNb9Xpx+GVLVc8k AHAM3BtACfeLuI
    JaXCctHTY0E/37c01pq+2cM=
    =ubrA
    -----END PGP SIGNATURE-----


+ Reply to Thread