Dell R510 with PERC 5/i - Foreign Config, lost VD

Hello

Did you pull a controller/TTY log? All of the information should be listed in that log. If you have restarted the system a lot then it may no longer contain the information. The controller/TTY log stores the last 10k lines.

If you don't want to have a data recovery company assist then I suggest using some type of drive cloning on the drives so that you have a backup. If you input any of the information incorrectly on a retag it will be data destructive.

http://www.dell.com/storagecontrollermanuals/

Thanks

Could you decipher this part of the controller log for me please?

T125: ARs configured: 1
T125: Ar 0: size=8780000, activeLd=1, numPd=4, seq=0, PDs: 00(08/0) 01(08/1) ff(08/2) 03(08/3)
T125: LD 0: L=0 SS=128 Size=21e00000 NL=3763 Status=0 DT=1655 BT=2614, prop=[ID=00,dcp=00,ccp=00,ap=0,dc=0,dbgi=0]
T125: span 0: sBlk=00000000, nBlk=08780000, ar=0

Is there anything else from the log file that is important to recovery?

Thank you!!!

-Rod

Could you decipher this part of the controller log for me please?

T125: ARs configured: 1
T125: Ar 0: size=8780000, activeLd=1, numPd=4, seq=0, PDs: 00(08/0) 01(08/1) ff(08/2) 03(08/3)
T125: LD 0: L=0 SS=128 Size=21e00000 NL=3763 Status=0 DT=1655 BT=2614, prop=[ID=00,dcp=00,ccp=00,ap=0,dc=0,dbgi=0]
T125: span 0: sBlk=00000000, nBlk=08780000, ar=0

Is there anything else from the log file that is important to recovery?

Thank you!!!

-Rod

Virtual disk 0 has 4 disks in it: 0,1,??,3

The ff on what is likely PD02 is an indication of a communication issue. PD02 is likely a faulty disk.

The logical disk is level 0, it is a RAID 0. It is a single span, so it is just a RAID 0, not a RAID 10 or some other hybrid array.

The stripe segment size of 128 is the default stripe element size of 64KB.

If you are retagging a virtual disk with a faulty drive then after you create the virtual disk you should offline the faulty disk to avoid it corrupting information.

Yes, the Virtual Disk 0 has 4 disks, 0, 1, 2 and 3.  Disk 2 is showing as missing in the VD Management under Physical Disks, yet it shows as "Ready" under PD Management as Drive 01:02.

I swear I set this up as a RAID5.  I obviously **bleep** up someplace if the log is reading as 0.  Even after accidentally clearing the Foreign Config, the VD Management showed RAID Level 'RAID-0'.  

What I do not understand, is why does the Config Utility for the controller still show RAID configuration info, if the foreign config was cleared?  I can send screen shots of the Config Utility if you like?  It shows the info that you replied with.  [ie. Stripe Element Size '64KB'...VD Size '271.000 GB']

What are my chances of recovery if I attempt a 'retag'?  I am almost %100 sure I have the correct configs, but I am just very hesitant to go forward with a retag, as I assume I will totally loose any possibility of data recover if it does not work?

Thank you!

-Rod

Data recovery gets more expensive with difficulty. The data may still be recoverable if the re-tag is not successful, but it may become very expensive to recover. All data recovery is expensive, but at this point it would probably be on the low end of cost. I suggest at least talking to a data recovery company before proceeding.

Since it is a RAID 0 with no redundancy you will need to bring all of the members online. PD02 entered a failed state at some point. The virtual disk would have failed as soon as PD02 failed, so there should not be much out of sync with the disks.

If you want to re-tag then recreate the virtual disk the same with all of the same members, don't offline PD02. Offlining faulty disks is not an option on a RAID 0.

Daniel,

Can you tell from the clip of the log file I posted what the VD size was?  I am going to attempt to retag, but I want to verify from the log file that the size is the same.  Would seeing the complete log file I pulled help?

I am leaving Advanced and Initialize unchecked...

Thank you, Rod

Size=21e00000
Convert from hex to decimal: 568328192
Multiply by the block size, I'm assuming your drives are 512. 568328192 * 512 = 290984034304
Convert from bytes to KB 290984034304 / 1024 = 284164096KB
Convert KB to MB 284164096 / 1024 = 277504
Convert MB to GB 277504 / 1024 = 271

271GB

Daniel,

I did end up retagging.  To my amazement, the server booted!  When I saw the black/yellow ESXI screen, I was quite happy!

Though, I ran into an issue.  My datastore is missing.  From everything that I can see, it shows the drive is blank?  After researching everything I can about this, I have come to two conclusions....

1) During the retagging, something did not 'match' correctly.  Either there are 2 partitions with ESXI, and the one that the OS runs from came back OK, but the one with the datastore, did not come back clean?

2) ESXI is not seeing the correct identifier for the drive that the datastore is on.  So the data is there, but the identifier is not seeing the data for some reason?

Thoughts?

Thank you,

-Rod

The retag process should have no affect on any partitioning. File systems and everything else installed on the virtual disk are just data to the controller. It is possible some data is corrupt. The only other possibility I can think of is that you had sliced arrays, multiple virtual disks across the same physical disks. If your drives are larger than ~73GB then you likely have sliced arrays.

I suggest that you get a backup before doing anything else. You should probably run diagnostics on the drives as well.

Daniel,

Sorry, I should also state that they were able to do this remotely, not with the drives in their facilities.  I made it so they could remote into the server after installing windows on a drive plugged into the On-Board SATA Controller in the PE1950.  They can see the RAID, but not the individual drives.  If I can get a different controller and setup using JBOD, then they could access the drives and rebuild the RAID outside of the config, as they would do at their facilities.

-Rod

A data recovery company should be able to see the raw data on the disk. If they are saying the data is truncated then they are saying that it does not exist. It is cut off at a certain point, and there is nothing past that point. There is nothing we can do to bring it back into existence. A data recovery company should have more options available than what I can suggest. If they say it is not recoverable then it is not recoverable.

Ok, I decided to hire a data recovery company to attempt to recover data.  They are able to get about %85 to %95 of the data.  They said that the reason that they can not get a complete recovery is because the RAID is Truncated?  As if the full size of the RAID was not entered into the config when I retagged the RAID?

Can this be fixed?

Thoughts?

Thank you,

Rod Mitchell