Re: Failed Software RAID-5
I got exactly the same problem here. The server was working ok, but it just freezed and became unavailable. I needed to manually shut-down (cutting the energy). On reboot md2 (RAID1) became unavailable with the same error as yours.
I have the schema:
/dev/md0 - /dev/sd(abcd)1 - RAID1 (boot)
/dev/md1 - /dev/sd(abcd)2 - RAID0 (swap)
/dev/md2 - /dev/sd(abcd)3 - RAID1 (/) *FAILS ON BOOT
/dev/md3 - /dev/sd(abcd)4 - RAID10 (file server)
When booting with live CD, all MD, except md2, start ok. I could even backup data from md3.
When looking for mdadm --detail /dev/md2 it shows the the spare as "active, not started". If I ask for details for each sd(abcd)3 some shows "active" and others "clear".
As if it is RAID1, if I mount sda3 manually it shows all the data. So it is still there in the drives. When trying to start the drive (even using --force) it says either "no superblock" or "mdadm failed to RUN_ARRAY /dev/md2: input/output error", but it mounts with the status shown above "(active, not started)".
It seems a common error in Linux (or opensuse), īcause itīs the second time it happens in less than 3 months. The other time it was a RAID5 and I lost all data. I though that with RAID1 I would be safer, but it is the one that is messed.
I hope anyone can help us.
Thanks,
|