S.M.A.R.T. Red light on system disk, test passed?

  • Hi,


    Just installed OMV 3.0 on a SSD instead of a normal HDD. Like the quicker feeling. But when checking the S.M.A.R.T. in OMV the status light is red, and if I hold over it, it displays:
    "Device has a few bad sectors".


    Did a test and this is the result:

    The result is:
    SMART overall-health self-assessment test result: PASSED


    Why is it showing a red light?

  • Okey, so something is wrong with the disk after all. It is an 80gb drive, which I am only using for OMV. Should I replace it or will it be fine as I am using so little of it?

    I use it until goes seriously bad, 2 bad sector is not so much, but can be a tendence.

  • Might be a good idea to make an image of it using clonezille.
    Is it a HDD? Why not exchange it by an SDD or USB Thumb drive. Might also save some energy.

    It is an SSD, 80gb Intel X25-M (older model).

    I use it until goes seriously bad, 2 bad sector is not so much, but can be a tendence.

    I will take a Clonezilla backup from time to time, will I get any warnings from OMV if more sectors fail?

  • will I get any warnings from OMV if more sectors fail?

    SSDs do not have sectors. The only SMART attribute worth a look is '233 Media_Wearout_Indicator' (with Intel SSDs -- other vendors, other attributes). This one goes from 100 down to 1 and is now at 99. As soon as it's at 1 you can take action based on SMART data (based on current usage this will be in a few decades or even +100 years). In other words: forget about SMART as health indicator, it won't work. Do backups and if I were you I would use also a checksummed fs and do regular scrubs (but on an OMV system partition a periodic 'dpkg --verify' should also do the job to alert silent data corruption)

    • Offizieller Beitrag

    on an OMV system partition a periodic 'dpkg --verify' should also do the job to alert silent data corruption

    What should the output look like?
    Something like this? If I understood google right the 5 between the ? means that the md5sum check is ok.


  • Might be a good idea to make an image of it using clonezille.
    Is it a HDD? Why not exchange it by an SDD or USB Thumb drive. Might also save some energy.

    It is an SSD, 80gb Intel X25-M (older model).

    I use it until goes seriously bad, 2 bad sector is not so much, but can be a tendence.

    I will take a Clonezilla backup from time to time, will I get any warnings from OMV if more sectors fail?


    Great, I will shut off SMART on the SSD system disk then.


    Did what you said. This is the result?
    Does it look good?


  • What should the output look like?

    Not suspicious ;)


    This command lists all files that have not the checksums as 'expected' (at the time the original package was installed). While some files are expected to change (eg. below /etc, the drivedb.h or everything with *cache* in its name) increasing occurences of system files are an indication for data corruption. As an example see here: Protocol "http" not supported or disabled in libcurl

Jetzt mitmachen!

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