[LUGSB] checking a software RAID partition for consistency

[Nicholas Lopez]

On Mon, May 5, 2008 at 12:08 AM, Mikhail Bautin <mbautin at gmail.com> wrote:
> but is there something like fsck for RAID that actually goes and checks the
> disks?

You can do a consistency / bad blocks check using:

echo check >> /sys/block/md4/md0/sync_action

and can monitor the check progress with:

watch cat /proc/mdstat

> Also, what is the standard monitoring tool for software RAID that would send
> me an
>  email when one of my disks crashes?  Any help is appreciated.

Most distros have an init script that stats mdadm up in --monitor mode
and sends the Critical level syslog events to an email address
specified in /etc/mdadm/mdadm.conf (these are Fail, FailSpare,
DegradedArray, DeviceDisappeared)

hth,
ngl