software raid issue

Hello everyone,

I'm having an issue with one of my machines where I had a software raid 1 ( I guess ) on two ADATA Premier Pro SP900, 256 GB SSD drives.
One of the drives died and my system began sending messages like this

Code:

lk_update_request: I/O error, dev sdb, sector 1400976

 sd 2:0:0:0: [sdb] tag#13 FAILED Result: hostbyte=DID_BAD_TARGET driverbyte=DRIVER_OK

The system was not usable anymore and I decided to reboot.

Result:
The system does not want to boot anymore and complains that it cannot find a lvm uuid.
When using a live cd I can see one disk id fd, Linux raid autodetect.
The second disk is dead.

I tried recovering using this:

Code:

fsck.ext4 -f -c -v /dev/sda1

but that said 'is in use' even though it was not mounted.

Assuming it was due to the raid I did the following:

Code:

mdadm --examine --scan

Who gave me one array back. It printed /dev/md/0 instead of /dev/md0 I see in the examples but apparently that should not be an issue.

Then I did an assemble

Code:

mdadm --assemble /dev/md/0

And a fsck on the resulting disk but it did not work either.

Then later I read that on a raid 1 you can just mount the disk as ext4 since it should be the same but that didn't work.

So my questions, what is the good way to recover from a software raid 1 failure and why did it affect my system in the first place?
Should I do an update grub and will my system boot then?
Should I have done the following instead of rebooting?

Code:

mdadm --manage --set-faulty /dev/md0 /dev/sdb1

Let's block ads! (Why?)

software raid issue

Read more ...