Hi!

I have a linear RAID array consisting of two drives/partitions, hde3 and hdg2. One of the drives went offline because of a faulty power-connector (fixed), but now the array won't start because they're out of sync (or something like that).
The events counters doesn't match, my logs say, and mdadm say hdg2 is dirty.
The array (md0) wasn't used (but it was mounted) at the moment hdg went down, so I don't think there's any real data-corruption.

Here's a peice of my log:
Aug 13 15:29:05 tomas kernel: md: linear personality registered as nr 1
Aug 13 15:29:05 tomas kernel: md: raid0 personality registered as nr 2
Aug 13 15:29:05 tomas kernel: md: raid1 personality registered as nr 3
Aug 13 15:29:05 tomas kernel: md: raid5 personality registered as nr 4
Aug 13 15:29:05 tomas kernel: raid5: measuring checksumming speed
Aug 13 15:29:05 tomas kernel: md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
Aug 13 15:29:05 tomas kernel: md: Autodetecting RAID arrays.
Aug 13 15:29:05 tomas kernel: [events: 00000075]
Aug 13 15:29:05 tomas kernel: [events: 00000011]
Aug 13 15:29:05 tomas kernel: md: autorun ...
Aug 13 15:29:05 tomas kernel: md: considering hdg2 ...
Aug 13 15:29:05 tomas kernel: md: adding hdg2 ...
Aug 13 15:29:05 tomas kernel: md: adding hde3 ...
Aug 13 15:29:05 tomas kernel: md: created md0
Aug 13 15:29:05 tomas kernel: md: bind<hde3,1>
Aug 13 15:29:05 tomas kernel: md: bind<hdg2,2>
Aug 13 15:29:05 tomas kernel: md: running: <hdg2><hde3>
Aug 13 15:29:05 tomas kernel: md: hdg2's event counter: 00000011
Aug 13 15:29:05 tomas kernel: md: hde3's event counter: 00000075
Aug 13 15:29:05 tomas kernel: md: freshest: hde3
Aug 13 15:29:05 tomas kernel: md: unbind<hdg2,1>
Aug 13 15:29:05 tomas kernel: md: export_rdev(hdg2)
Aug 13 15:29:05 tomas kernel: md0: max total readahead window set to 124k
Aug 13 15:29:05 tomas kernel: md0: 1 data-disks, max readahead per data-disk: 124k
Aug 13 15:29:05 tomas kernel: md: md0 stopped.
Aug 13 15:29:05 tomas kernel: md: unbind<hde3,0>
Aug 13 15:29:05 tomas kernel: md: export_rdev(hde3)
Aug 13 15:29:05 tomas kernel: md: ... autorun DONE.

So, the question is, how do I fix my array?

Thanks in advance!

Ok, I've fixed it.. I'm replying to myself incase someone else with this problem comes across this post.


First, I thought, if I recreate the RAID-array EXACTLY as it was, the data on the old one shouldn't be erased, right? mkraid needed the /etc/raidtab, which I of cource no longer had, so I executed

to recreate it, then ran

to create a new array (on top of the old one). The -R option is to force mkraid to create the array, even though the partitions appears to contain data.

Now the array was up again, but it wouldn't mount , so I ran

to fix it.

I don't know if it'll work for anyone else, but at least, if you're desperate enough, you could try it as a last resort before reformatting the whole thing...


My array is back