lost my raid... urgent!

[VJ]

Hello,

It is turning out to be a bad evening: I was happily ripping my cd's and dvd's (almost finished) when my computer just froze.
Upon the reboot the non-raid dives refused to spin up, so I openend up the case pressed the powercables and retried - with succes (it has been very hot, and I've seen this happen on my system). But now I saw a raid bios message saying that there was a drive failure (raid was in a critical state. Apparently, I knocked out its SATA cable. So I plugged it back in, and started rebuiling the raid. About 18 percent in the process the raid went offline!

Dunno why, but now it lost another disk in the raid. I did notice that the sata cables are quite loose on the controller side :-(. Rebooting now, and it sees a raid 5 with one disk missing and a free drive. And I can't seem to tell it that this one disk is part of the raid. And as I can't do that, I fear I may have lost all data stored on it.
The controller is a Promise sx4300, all drives appear healthy.
I see no alternative but to delete the logical drive and recreate it, but I assume this will delete all data...?
Any other things I could try to regain my data?

Thanks!

Jorg

[Gurm]

Sounds like you've gotten it good and confused. Other than unplugging all the drives, plugging them in one at a time in the correct order? Don't know what to tell you. It's pretty much decided that drive is no longer part of the RAID, so it's probably borked at this point.

[UtwigMU]

Probably the drives are OK. If not, you can try rescuing damaged drive by sticking it in another machine booting Knoppix and doing dd_rescue to empty HDD

You could try getting another Promise controller and importing the RAID there.

There are also services that do RAID rescue such as Ontrack Germany. It will be more than 1000 Euros for a RAID array though.

[VJ]

All the drives are ok, that I can see. The first problem was caused by a loose SATA connector; I suspect the second problem as well (all these SATA cables don't click well into place ).
The biggest issue I have is that the system doesn't see that all drives form a single raid. It knows that disks 2 and 3 belong to it, but due to the first problem disk 2 was in the process of a rebuild. Disk 1 is seen as free drive, not as a drive participating in the raid.

Importing a raid... I wonder if I could import the raid back in my current system...? I'll have to check on how to do that...

(The irony is that if the disks would not have been in a raid setup, there would have been no problem )


Data recovery is not worth it, I lost most of the free software I use on my PCs (different OS), mobiles and Psion. This can be downloaded again, main problem is remebering and finding them (some softwares are old and not maintained anymore, others were older versions that I prefered to the newer versions but that are no longer available). The worst is that I had nearly finished ripping all my CDs to flac (and tagging the files completely: with publishing info and everything), and now I'll have to redo that (takes about 4 weeks).

Normally, everything is backed up, but my backup disk was full, and I was cleaning it (contained lots of old data that I could remove from it).


Jörg

[az]

For help tagging, try MusicBrainz Picard. It really makes the process much easier.

[VJ]

Problem isn't with the regular title/artist tags, these are found/entered easily enough. But I wanted consistency in capitalization and was adding additional information (producer, people involved, track information, ...); this information isn't automatically filled in.

Luckily I hadn't started with scanning the booklets...


Jörg

[VJ]

A forum member pm'd me with a solution:
1. delete the logical drive
2. redefine the array

I called Promise tech support and they confirmed it, adding that it is imperative not to initialiase (fast initialize=rewrite mbr) the newly defined array.
There is a warning message that pops up when you delete a logical drive (stating that it will erase all data), but the tech support confirmed that no actions are undertaken that cause deletion of data.
Naturally, they cannot guarantee anything, but it is just the only procedure to try and recover data.

AND IT WORKED!!!!!

My raid is back up and running (rebuilding from that first failed disk), and now I'm copying all relevant data to a backup disk.

Woohoo!!!

edit: given the cache settings, I a couple of files did go corrupt. In total I think I might have lost 20 or so files. So no big deal.

Jörg

[The PIT]

Backing up after the event LOL.

[VJ]

Backing up after the event LOL.

Better late than never...

My backup disk was too small, so at the time of the crash I was re-organizing it. A typical Murphy's law event...

The lessons are:
1. use sata cables with latches
2. improve organisation on backup disk (easier maintenance)


Jörg