I've got a linux software raid system set up. It is raid 5 across 4 disks.
We recently started using smartmontools and I have found that two disks
in the raidset have failed an extended offline test.
# 1 Extended offline Completed: read failure 10% 12581
# 1 Extended offline Completed: read failure 10% 12518
hdg has one error logged but from ages ago, while hdi has no errors logged.
Unless I'm very unlucky the unreadable sectors on both disks aren't in
the same file so the raid system can reconstruct the data correctly at
However, I'd like some advice on finding any files affected, and on
forcing the bad sectors to be reallocated please. I've read the badblock
howto but I don't think its really applicable to the raid 5 situation.
If I only had errors on one disk then I could just fail it and
reconstruct the array, but this will cause a problem in the files that
have bad blocks on the other drive.
Any good ideas before I backup reconfigure and restore this large raidset?