[SGVLUG] software raid check -- was Re: Moving files in a safe way

Claude Felizardo cafelizardo at gmail.com
Sun May 3 19:38:42 PDT 2009


I've had various problems where my software raid either didn't start
up correctly or a drive got an error and was marked as bad.   I now
run a periodic check via a cron:

30 2 * * Wed               echo check > /sys/block/md0/md/sync_action
30 2 * * Wed               echo check > /sys/block/md1/md/sync_action
30 2 * * Wed               echo check > /sys/block/md2/md/sync_action
30 2 * * Wed               echo check > /sys/block/md3/md/sync_action

The checks on the small raids complete fairly quickly but takes hours
on my big raid-5 drive.


On Fri, May 1, 2009 at 1:15 PM, Matt Campbell <dvdmatt at gmail.com> wrote:
> Thanks Ted,
>
> It's a trusted network (my house, my daughter isn't quite six yet so I'm not
> running across domestic terrorism issues yet).
>
> The target/source machines are Intel/Fedora boxes running largish RAIDS.
> The source machine is FC6 with two 4T software(md) arrays and the
> destination machine, Fedora 10, has a 10T hardware(Adaptec) array.
>
> Matt


More information about the SGVLUG mailing list