[LUNI] oh oh, my raid it having problems

[Jay Strauss]

On 8/2/07, Craig Van Tassle <craig at codestorm.org> wrote:
> I would check /proc/mdstat and if it shows one drive failed then I would replace the drive.
> You should be able to tell what drive is bad when you look at mdstat. I'm assuming you are only using the full drive for
> raid. Get a drive of the same size, remove the failed drive. Install the new drive. Partision it and then hotadd it to
> the array. Depending on size of the partision you should be able to sync it from any time between 10 minutes to 3 hours.
> watch -n 1 'cat /proc/mdstat' will show you the drive as it resyncs. It should not take that long.
>
> Then I would check RMA status and also run a badblocks on the failed drive. That should tell you what went wrong.
> Another thing to do is check the smart status of the drive. That should show if the drive failed and why.
>
> HTH
>
> Craig

Hi Craig, thanks for advice.  Here is the output:

iron:~# cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 hda1[0]
      1951744 blocks [2/1] [U_]

md1 : active raid1 hda2[0]
      115266304 blocks [2/1] [U_]

unused devices: <none>

I'm guessing it's hda1 that is the problem.  I don't know how to print
the partition table.  I have 2 disks which I broke into 2 partitions.
On top of those I built two raid 1 devices.  On one of the devices
(the big one) I built an LVM.

So how do I print the partition table?

Do I power down the box, remove the drive, replace with new, power
back on and try to sync OR do I do it all while the box is running?

What is RMA how do I check it?  Do I stick the failed drive in a
different computer?

Thanks
Jay