Il 13/09/2012 12:34, Robin Hill ha scritto:
Has anything actually attempted to read from that part of the array?
Even if so, it may just have happened to read from the working disk
anyway. md can only detect the error when it tries to read/write that
sector of that disk.
I forced a read with "dd if=/dev/md0 of=/dev/null bs=524228 count=1
skip=1", I even get errors in syslog!
Your best bet now is to do an array check:
echo check> /sys/block/md0/md/sync_action
This will force a read of all disks in the array. This should trigger
the read error, causing an attempt to re-write the faulty block, in turn
causing the drive remap the bad sector (assuming the re-write fails).
This should also be scheduled to run regularly for all arrays in order
to pick up these sort of issues before they cause major problems during
a rebuild.
/etc/init.d/mdadm should do exactly this kind of things (distro is
Debian Squeeze). I have this in cron.d:
57 0 * * 0 root if [ -x /usr/share/mdadm/checkarray ] && [ $(date +\%d)
-le 7 ]; then /usr/share/mdadm/checkarray --cron --all --idle --quiet; fi
Unfortunately it seems it didn't work :(
Shouldn't a dd if=/dev/md0 be enough to trigger the read error?
Thanks,
Niccolò
--
http://www.linuxsystems.it
--
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
