2 out of 4 disks defunct in RAID5 set

[Laggetjeuh]

Last night during a power outage one of the disks in an old server of ours decided to go to defunct status.
Unfortunately this server originally had 4 disks (1 logical drive) in RAID 5 and about 2 years ago one of the disks already failed, which left us with 3 disks (the minimum for RAID5) then.

Somehow back then it slipped our minds to order a new disk for the defunct one and today we´ve met the consequences... Another disk failure and the server won´t boot anymore.

As usual in these cases the backup job hasn´t been running well for the last couple of days so if we restore all files from tape the users will have lost over 2 days of work

Does anyone know a way to get the drive that failed to go back online so we can access the data again? (I know theoretically it´s impossible because you need at least 3 working drives in this configuration)

We´ve tried setting the drive back to online using serveRAID, updated everything to latest version possible for this kind of hardware(IBM Netfinity 5100 with 9GB SCSI disks and a ServeRAID 4L controller), used IPSSend to try to set the disk back online but all of this failed...

I hope someone can help me out here... Kind of out of options right now.

Thanks in Advance!

Jeroen.

 

[ProbablyDown]

Same thing happened to me... (except I lost 2 at the same time). You'll probably have to do what I did, which is go to a data recovery company...

It's gonna cost ALOT more than that defunct disk would have... :-(

Sorry... but I think you're effed...seriously though, it cost me around CDN$4000, with quotes as high as CDN$10,000 - but the data was worth alot more, so...

 

[ProbablyDown]

By the way, I think if you lose 2 disks in a raid5 you're out of luck no matter how many disks are in the array...

 

[ibmtech65]

If you boot to servraid manager support CD and right click on the drive there should be an option to force drive online. If that does not work than you are toast.

You said you updated the firmware??? do you remember what was the original version and what is the new version because you now might have a problem with either the Metadata on the drive config and/or driver issue.

As a last ditch effort, provided the controller can still read the drives, you can right click on the controller and restore to factory defaults. If you then get 4 Ready drives you can do what is called create the array around the data.
You would go through the steps to create a new logical drive but making sure that you UNCHECK the check box that says automatically initialize the logical drive. If you do not do this it will erase everything.

If you had just the one logical drive then add the drives and create a raid 5 array. BUT before you click on apply I would physically remove the original failed drive from the machine then click on apply. You will probably get some warnings and stuff like that but dont worry since at this point you have theoretically lost everything anyway.

If that works then you should be able to boot up to the OS provided that you dont have a major mismatch between the firmware and the driver. If it does not work then you are right back were you started so nothing lost since it is already lost.

 

[Laggetjeuh]

Well the drive won't go online, no matter what I do.
If I would recreate the raid array would this do something to the 2 disks that are sdtill functioning?

I just don't want to lose these 2 disks since we may need em if it really comes to getting the defunct disk repaired by a company...

 

[Laggetjeuh]

I´ve also tried the 3 disks in a similar server and let the controller get the config from the drives, however this ended with the same result

 

[ibmtech65]

Those other 2 drives are really not doing anything. You lost 2 of your data drives. If anyone was to attempt to try to get the data off they would need atleast 3 good working drives. If that other drive will not allow you to set it online then there is a reason for that and it is probably not going to work.

The config on the drives is always going to show 2 DDD and 2 ONL no matter if you move it from machine to machine.

Best thing to do is get some new drives or new server and reload from backup. 2 days of work is a lose but it could have been worse. At least now there is a good reason to push for a more reliable backup system.

 

[ProbablyDown]

Like I said - if the data is worth something, take the drives (All of them) to a data recovery place and they'll toss all the data on to on working drive...

 

[Laggetjeuh]

done that... hope they can get smth off the disks...
thanks for your help all!