Recovering old Raid6 Array: LSI 8708 EM2

[whiskytangofoxconn]

It's been a while and I'm looking for a sanity check.

Can someone educate me on lsi 8708-em2 raid cards? I've got an old raid6 array attached to an 8708-em2 card and two of the drives are bad. The underlying drives in the array are from the bad batches of Seagate drives way back that were notorious for being bad. I'm building a NAS and need to buy drives. Ideally I'd buy a bunch of big spinny drives, do a dd copy of the 2TB drives over to the bigger drives, recover and grab the data off the raid6 array, then put the bigger drives in the nas and be done with it.

But, if I remember correctly, didn't that 8708-em2 card have a 2tb drive size max capacity? I have a couple lsi 9260-8i cards. Is it stupid to try and import an already degraded array off of the 8708-em2 card and onto a 9260-8i card that can handle higher capacity drive sizes? It is a 7tb raid array so I don't really want to risk copying 6+ TB of data from a degraded array. If the 8708-em2 card doesn't play nice with drives over 2tb, are there any other options to move the degraded array elsewhere?

Thanks.

The machine has been sitting powered off for a few years and I have no idea what is on the array. It could be nothing (duplicates of things I have) or it could be data I don't want to lose.

 

[Samir]

If I were in your shoes facing this task, the first thing I would do is physically clone all the drives to another 2TB drive. This will give you a working set from the originals. Then dd these copies so that you have the drive data safe. Then I would literally just try booting the raid and copying off the data. There is less of a likelihood for an original drive to fail since they're not being touched, and you can recreate the entire set of drives if need be since you have a dd copy of each drive.

Then once you're done recovering, you've got all these drives to do with as you wish.

 

[whiskytangofoxconn]

If I were in your shoes facing this task, the first thing I would do is physically clone all the drives to another 2TB drive. This will give you a working set from the originals. Then dd these copies so that you have the drive data safe. Then I would literally just try booting the raid and copying off the data. There is less of a likelihood for an original drive to fail since they're not being touched, and you can recreate the entire set of drives if need be since you have a dd copy of each drive.

Then once you're done recovering, you've got all these drives to do with as you wish.

Thanks Samir, I ended up buying a few 2tb hdd's. When they get here I'll do as you suggested and clone (dd) the drives then I'll start copying data off the array. As I thought, that hw raid card won't talk to drives over 2tb. I needed a few larger size drives and thought I could dd copy the 2tb drives over to 6tb drives but the hw raid card won't talk to the 6tb drives.

Thanks again.

 

[thetoad]

use ddrescue, in case you are missing a handful of blocks, you can still try to recover data, though some will be corrupted.

 

[Samir]

As I thought, that hw raid card won't talk to drives over 2tb. I needed a few larger size drives and thought I could dd copy the 2tb drives over to 6tb drives but the hw raid card won't talk to the 6tb drives.

Yeah, I've run into the 2TB limit before as well. Best best is to simply get some 2TB drives. And if you need help sourcing some good ones, let me know as I have seen some lately.

 

[thetoad]

if its only a ~7TB array of 2TB discks in raid6, I'd assume that means a 6 disk array with 2 parity drives?

if I were in your shoes, I'd buy a 12TB disk or larger (or 2). ddrescue each individual drive to the disk. use a tools like rstudio to rebuild the raid array from the images and use it to copy out all data (hence the possible need for the second disk)

 

[pricklypunter]

I would not waste my money on 2TB disks now. Buy some decent sized disks, build your new array with the LSI9260-8i and create whatever partition scheme you like. In addition also create a "rescue" partition, then dd/ ddrescue the 2TB disks to the new "rescue" partition. Rebuild your old array from the image files and recover your data to where you need it. Once you get back your critical data, you can either keep the image files or blow them and the rescue partition away, but I suggest only after a few days have passed and you are certain you haven't missed anything

 

[whiskytangofoxconn]

Thanks for all of the replies. I ended up freeing up 6x spare drives from a server I had mothballed a few years back. I thought it had written down that it had 3x 5tb drives in it (mirrored with spare) and it really had 6x 2tb drives mirrored & striped.

I got everything "rebuilt" only to have Windows server 2008 (Datacenter) tell me that the healthy partition was empty. That was fun.

Next I swapped out the OS boot drives and installed Ubuntu so I can run disk utility tools. Scalpel has been running and pulling files/data for a few days. The only good thing is that once this mess of 2tb hdd is finished I can throw out those 2tb drives and be done with every last Seagate 2tb disaster drive I own.