Check Hard drive for bad sectors or bad blocks in linux

7 Responses

  1. David Wood says:

    don’t you need to make sure the blocks are not allocated to a file or swap etc?

  2. Bernardo Verda says:

    Don’t modern harddrives already have all this already — as built in functionality?

    (access via smartctl)

    $ apropos smartctl
    smartctl (8) – Control and Monitor Utility for SMART Disks

  3. Виктор says:

    From time to time, it is a good practice to check your hard drive for errors using Windows CHKDSK or Check Disk.

  4. deco says:

    This has given me an idea of creating this script

    for disc in `fdisk -l | grep ‘^/’ | awk ‘{ print $1 }’`; do
    badblocks -v $disc > “$target”
    tam=$(du -k “$target” | cut -f 1)
    if [ $tam -eq $minsize ]; then
    echo “no badblocks on $disc”
    echo “badblock(s) found(s) on $disc”

  5. cipher says:

    You don’t need to run badblocks first. That’s a waste of time. Just use the e2fsck executable to run it for you with the -c option…

    # sudo e2fsck -c /dev/sdb

    This lets you avoid creating an unnecessary file, especially when you’re booting to an optical drive and don’t want to write to the hard drive you’re testing.

  6. Stoat says:


    The ONLY reason you should ever create a bad blocks table is to send commands to the drivbe to tell it to repair the sectors

    ATA and scsi drives are supposed to map out bad blocks. You can force the issue by wirting 0x00 to the sector (or use hdparm–repair-sector)

    Take note of smartctl -A /dev/drive returns – in particular the bad sector (mapped out sectors) and pending sectors (probably bad but not mapped out) values – if the bad sector normalised value (not the raw value) is below the threshold then you need to replace the drive – NOW.

    If you have large numbers of pending bad sectors then you’ll need to identify them and write zeros to them to allow the drive to fix or map them out (perfectly sectors can get marked as pending due to vibration issues, amongst other things).

    Badblocks -svn will do this (unmount the partition first!), or if the partition is trashable/part of a raidset, hdparm –repair-sector and then rescan/reformat afterwards.

    If bad sectors are not being mapped out then the drive needs to be replaced immediately. It’s probably out of spares (you’ll see this with smartctl -A) or about to fail.

  7. techhead says:

    What command just to report sectors or drive health.
    Using an old pata drive that has not been used in some time to run Linux
    peppermint 64x.
    From my experience with windows os’s and smart, one bad sector can trash
    a program or even the entire drive.

Leave a Reply

Your email address will not be published. Required fields are marked *