One could remove the spare drive from the system. Than do the mdadm --assemble --force to get it start and keep it from trying to resync/recover. Once you get the array up and 'limping' carefully pick the most important stuff and copy it off the array and hope the bad sectors don't did not affect that data. As you mentioned, you have bad sectors. If you try to do it all or even as you pick the important stuff (I know it is all important that why it was on the RAID), you will eventually hit data that has bad sectors and mdadm will fail the effected drive and deactivate the array. At that point accept that the data in that area is most likely gone. Than do the mdadm --assemble --force to get it started again and move on to the next areas of most important data. Could be a long cycle... Aside from that, I am curious, was your spare disk shared by another array. If it was not than I would recommend you don't do a RAID5 with hot spare next time, and do a RAID6. But this is my person feeling and you can take it or leave it. I too at one point did a RAID5 with hot spare and I was using eight drives. So yes you can have two drives fall as long as the delta between failures is long enough to allow the raid to resync the spare in and during the process there are no be unknown bad sectors on the remaining drives. But I got to thinking, if I am going to be spinning/powering that "hot spare" I may as well as do a RAID6. As long as the hot spare is not share with other arrays, I see no downside and it would protect you in the future from this problem. -----Original Message----- From: linux-raid-owner@xxxxxxxxxxxxxxx [mailto:linux-raid-owner@xxxxxxxxxxxxxxx] On Behalf Of Lane Brooks Sent: Saturday, October 14, 2006 9:29 PM To: linux-raid@xxxxxxxxxxxxxxx Subject: Multiple Disk Failure Recovery I have a RAID 5 setup with one spare disk. One of the disks failed, so I replaced it, and while it was recovering, it found some bad sectors on another drive (unreadable and uncorrectable SMART errors). This generated a fail event and shut down the array. When I restart the array with the force command, it starts the recovery process and dies when it gets to the bad sectors, so I am stuck in an infinite loop. I am wondering if there is a way to cut my losses with these bad sectors and have it recover what it can so that I can get my raid array back to functioning. Right now I cannot get a spare disk recovery to finish because these bad sectors. Is there a way to force as much recovery as possible so that I can replace this newly faulty drive? Thanks Lane - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html - To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
