Repairing Offline JBOD RAID disks

[etl]

I am trying to undue an extremely stupid mistake I did on a user's Mac Pro machine.

The machine needed a firmware update. It had 2 500GB drives concatentated.

I realized that the firmware could not be installed booting from a RAID array. Thinking (not the best choice of verb in retrospect) I would split the RAID, update the firmware, and reconnect the RAID...I used the destroyRAID command. (it still hurts to type that) I have done no other manipulating since but boot from the install DVD and boot it into target disk mode.

The user had about 100GB of data. As I understand JBOD and Apple's implementation of concatenate, there is no striping. So all of the data should be located on the first disk. The software raid info is stored on the drives so I am hoping it is still available and I can use it to restore the raid.

Here is the output from diskutil check and list when the mac pro is in target disk mode attached to my workstation:

RAID SETS
---------
No RAID sets found

/dev/disk3
#: type name size identifier
0: GUID_partition_scheme *465.8 GB disk3
1: EFI 200.0 MB disk3s1
2: Apple_RAID_Offline 465.4 GB disk3s2
3: Apple_Boot 128.0 MB disk3s3
/dev/disk4
#: type name size identifier
0: GUID_partition_scheme *465.8 GB disk4
1: EFI 200.0 MB disk4s1
2: Apple_RAID_Offline 465.4 GB disk4s2
3: Apple_Boot 128.0 MB disk4s3


Have searched all types of boards most of the night for an answer but have come up dry.

Tom

 

[sourcehound]

The user had about 100GB of data. As I understand JBOD and Apple's implementation of concatenate, there is no striping. So all of the data should be located on the first disk. The software raid info is stored on the drives so I am hoping it is still available and I can use it to restore the raid.

Tom, while it's true that the concatenated RAID set does not employ striping, it's still a single file system volume, meaning that you will lose all of the directory information of you break up the concatenated set. Because there's no striping, you should have litle troble recovering the data via Data Rescue or File Salvage, but the volume and catalog and partition information are permanently gone.

It's not really surprising you misunderstood the difference between concatenated and striped RAID. There's little documentation about it. But here's the basics:

1. Conecatenated RAID is designed to make volumes of ***different*** sizes appear as a single volume, with performance comparable to a single volume. Volume information is held on a special structure on all of the disks, without copies on each. JBOD is safer than this configuration. In a data recovery situation, whole files ARE recoverable from any of the disks in a concatenated RAID set.

2. Striped RAID is designed to make volumes of the same size appear as a single volume, with enhanced performance over a single volume. Volume information is held on a special structure on all of the disks, without copies on each. JBOD is safer than this configuration. In a data recovery situation, whole files are NOT recoverable from any of the disks in a striped RAID set.

 

[swensenj]

I have a similar problem. I used DiskUtils to create a Striped Raid and in the process of making a repair clicked the wrong button. Now there is only one of the two volumes in the raid definition. Is there a config file or some other way to re-designate the second drive to be part of the raid volume?