I've got a system with 4 disks set up as raid10. All disks are usable, and mdadm all states them with UUUU.

Due to a recent system crash, the raid is currently reconstruction the raid as it was marked as "not clean," and a reconstruction process was started.

On a closer look smartctl shows problems on one disk:

sd 0:0:0:0: [sda] Unhandled sense code
sd 0:0:0:0: [sda] Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
 sd 0:0:0:0: [sda] Sense Key : Medium Error [current] [descriptor]
Descriptor sense data with sense descriptors (in hex):
        72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 
        24 cd 78 d4 
sd 0:0:0:0: [sda] Add. Sense: Unrecovered read error - auto reallocate failed
sd 0:0:0:0: [sda] CDB: Read(10): 28 00 24 cd 75 1e 00 04 00 00

With a research about the reconstruction process, I only found information concerning raid5 but nothing for raid10. Can I replace this problematic disk during the reconstruction process, or will I kill the raid with this?

migrated from stackoverflow.com Oct 29 '12 at 16:58

This question came from our site for professional and enthusiast programmers.


It would be best if you remove the hdd from the RAID array with mdadm, power off the machine, change the hdd, turn on the machine and add the new hdd to the current array. This should not destroy the reconstruction.

I don't know about others, but I had some bad experience with sw raid1 and bad disks. As far as I have found out, sw raid1 doesn't have ECC and will not check data for consistency. As I am unsure of this information and it is possibly wrong, I did not want to play with it since this were production machines.

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.