ZFS volume share by accident mounted four drives instead of the six: FreeNAS v9.1

[pagetelegram]

Back in 2012 we setup a FreeNAS raid-5 for 6 drives mostly for our video storage and source video editing over gb connections.

I just tried reviving the server however before the BIOS was setup to recognize all 6 drives initially loaded just 4 that I mistakenly imported as a single volume.

Now I am unable to get all six to mount. With four mounted the volume reported it as empty.

 

[garm]

If you used the four drives to build a new pool the old data will be gone. Please describe precisely what you did, terminology matters.

 

[pagetelegram]

four drives imported as one volume instead of the 6. Usually takes eons for zeros to write over data and unused space so I would think this scenario would be recoverable since no new data has been written to the drives (at least not by myself.)

 

[garm]

ZFS will not import a pool with missing drives, so you didn’t import the pool. Please tell us exactly what you did

 

[pagetelegram]

I'll report back when I arrive at the office where the FreeNAS presides.

 

[HoneyBadger]

Back in 2012 we setup a FreeNAS raid-5 for 6 drives

Assuming you intended this to be a RAIDZ1, if you have four out of six drives present, you don't have enough drives to have a functional pool, which might have explain why the mount point is empty.

Please describe the hardware in the server, with particular emphasis on the way the drives are connected and if any additional interface cards (HBAs) are in play.

 

[pagetelegram]

Using Supermicro Xeon server slate, 6 sata connections. Using plop CD to boot the USB that contains the FreeNAS.

Here are some screenshots:

 

[Fredda]

Please go to the shell and post the output of zpool status

 

[pagetelegram]

zpools:

 

[garm]

Using Supermicro Xeon server slate, 6 sata connections. Using plop CD to boot the USB that contains the FreeNAS.

Here are some screenshots:

Did you use the volume manager when you attempted the pool import?

 

[garm]

zpools:

Okey.. you definitely do not have a four drive pool. You have a single drive pool with three spares.

My money is on you lost your data. Maybe someone will come along and suggest a recovery route.

Just for good measure, what does zpool import say?

 

[pagetelegram]

So is it possible I had encryption? Too long ago to recall.

 

[garm]

No it’s not encrypted, you should just have followed the manual for pool import instead of nuking the drives. zpool import sees your remaining drives and want you to supply the missing drives, those you used to create a new pool

 

[pagetelegram]

Confused. Is there a possible recovery path at this point?

I did a volume import initialy only when the system recognized just four of the six hard drives.

 

[garm]

a data recovery company might be able to dig out the old pool. There is no ZFS toolset or consumer available third party toolset that I am aware of that would be able to reconstruct the pool.

If you truly did import your pool then someone else already nuked it for you. Someone has run the volume manager to create the single drive pool with four spares. Sins zpool import can set your original pool those drives are obviously intact and at this point that is the only bit of good news, not that I know what it will benefit.

Restore from backup is my suggestion, but I guess that isn’t an option?

 

[pagetelegram]

No, backup restore not an option.

A simple search revealed a $300/yr tool as an option. Though I don't have any Windows system currently to test it out with exception of some old Win NT4.0 machine. And $300 excessive amount to pay for our non-existent budget.

This server was used for a NFP organization and space before the area we lived in got gentrified. New landlords=more rent.

 

[danb35]

You have a single drive pool with three spares.

It's amazing the ways people come up with to destroy their data. And how little attention they pay to the message that says all data will be destroyed, until after it's already happened.

@pagetelegram, unfortunately, the data on this pool is lost, and it's through a series of very poor decisions:

• This was set up as a RAIDZ1 pool, which only provides one disk's worth of redundancy
• Rather than the six-disk RAIDZ1 pool that was intended, it was set up as a five-disk RAIDZ1. Apparently recognizing this error, they added the sixth disk (ada4) to the pool. But that disk was never part of the RAID set, meaning that it was always critical--failure of that disk alone at any time would have resulted in loss of the pool. It was also apparently added from the CLI rather than the GUI.
• Bad as all that was, it wasn't enough to result in data loss--that didn't happen until you created a new pool on four of the six disks, and in a truly bizarre configuration. When you did that, you apparently ignored the bold red text here:
  
  Do I know that you did this rather than someone else? No, but it sounds like it. You can check for yourself by looking at the output of zpool history mk--that will tell you when the pool was created, and by extension, who did it.