0

I have a server with a 3Ware 9650SE 8 port controller. I had an 8 drive raid 5 array with a hot spare drive as part of the array of roughly 3.87 TB in size. When I signed into the server today, I found that I have now TWO arrays - 1 array with all drives ok, but the hot spare is no longer a member, and a second array with only the host spare in the array (same size) with only the hot spare marked as good and all the other drives (which are OK members in the other array) marked as degraded.

I have a backup, but its uncertain if its corrupted.

Any advice on what caused this to happen. What to do next? should I delete the second array and merge the drive back into the functioning array?

2
  • 1
    Have you contacted 3Ware support? That would be my recommendation.
    – joeqwerty
    Feb 11, 2010 at 3:15
  • Are you sure, that the second array includes the disk you originally designated as the spare device? I remember something like this, where the new array with only one disk was the device removed from the original array (and it had of course all remaining disks missing). So check your logs if the array was rebuilt in the past and always use event-notification when possible.
    – r_3
    Jan 24, 2015 at 18:52

2 Answers 2

2

Pretty strange (to say the least).

It sounds like someone (or something) has cause your RAID volume to split. If anyone else has admin authority over this device, I'd ask them if they might have inadvertently reconfigured the volume (maybe they thought they were looking when in fact they were changing). In any event, your primary RAID volume is in degraded mode because it no longer has a hot-spare.

I'd first make absolutely certain I have a 100% good backup (preferably a bare metal backup). The try to re-add the spare to primary RAID volume (the 3.87 TB) volume. You will probably have to delete the goofey new volume before the 3Ware controller will allow this operation.

WARNING: 3.87 TB is big and it will likely take you several days to recover from tape if you need to rebuild your primary RAID volume. If you can wait until Friday @ 5pm (assuming you're organization is not 24x7) and do the work, that would be the best choice. This way if you are forced to recover from backups, you'll have the weekend.

Jim C. http://www.kleobackup.net

1
  • As far as we know no admin caused it. 3ware support is equally puzzled and took the logs for further study. We have a hot backup on one one those WD sharespace units. we are going to pull down the server tomorrow becuase its a holiday in canada. replace the hot swap drive and merge it back in. if that doesnt recover with the verify cycle, its delete everything and restore from the hot backup.
    – MikeJ
    Feb 14, 2010 at 16:44
1

[5 years later] I have the same 3ware card. This past month, I've had a drive not come online a few times. This causes the normal need to remove/rebuild (logically, not physically) the RAID5 array.

However.. ..twice in the last year, the troublesome drive comes online thinking IT is the RAID5 array and everyone else is missing-- split brain. Delete the extra RAID5 and rebuild the original (well in your case, make it the spare because it is already rebuilt).

It indicates that one or more drives are not powering on correctly. Either they're the wrong kind of drive, there is a power issue, or the drive is failing in a wait that does not otherwise report errors.

It is a pity that competent support is no longer available.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .