Installed FreeNAS 11RC3 and wizard failed to import existing volume

[shpxnvz]

Hi, I installed Corral a while ago and it worked fine for a long time. Recently I've been having network issues so I wanted to update to the new supported version. With FreeNAS 11 on RC3 I decided to give that a shot. I followed the instructions at https://doc.freenas.org/11/install.html (which was basically to install over top of the existing Corral install.

Once it finished I went to the website and ran the Wizard as the intructions said, which showed my existing volume to autoimport. When I finished the Wizard it ran for a while then stopped and popped up a window that said it failed but had not more detail than that and the UI was unresponsive.

When I log back into the web console it comes up but no Volumes are found. If I try to run the Wizard again it no longer sees my old Volume and instead asks if I want to import (and erase) all my disks, so I exited out.

I'm pretty lost and not sure what to do to see if my data is still there, and if it is how to add the existing volume now?

My hardware/configuration is outlined in my old install thread, I'll look up the link and update in a few.

Thanks in advance for any help.

ETA: My hardware setup - https://forums.freenas.org/index.php?threads/new-installation-on-existing-hardware.50442/

 

[Alecmascot]

Did you follow this : https://forums.freenas.org/index.ph...ng-from-freenas-corral-to-freenas-9-10-11.36/

 

[shpxnvz]

Yep that's what I started with. In this section:

How do I move to FreeNAS 11?
If you have a previous 9.10 boot environment, boot that up and switch trains to the FreeNAS-11-Release train. You will receive an update to the latest version.
If not or if your 9.10 environment is acting up, just install the latest version from ISO. It is available at http://download.freenas.org/11/

Since I did not have a previous 9.10 install I did the fresh install from ISO as directed.

 

[nightshade00013]

It's better to just repost your specs than to have people going to another thread.

Was the pool encrypted? If so your data may be toast unless they fixed the error. And since it failed it may be toast anyway as when the import would fail it would delete the pool.

There is a reason why it's not a wise idea to run beta or release candidate software on a production machine.

https://forums.freenas.org/index.php?threads/pool-loss-bug-11-1-rc1.59338/

 

[shpxnvz]

No not encrypted. He's the specs:

Motherboard:
Supermicro X8DT6

CPUs:
Dual Xeon X5650

Memory:
48GB Micron 36KSZF1G72PZ-1G4D1 DIMMs

Disks:
12x 2TB Seagate Constellation ES ST2000NM0011

Controller:
LSI SAS2008 SAS-2 controller, flashed to pass-through mode (don't remember which firmware version at the moment)

Other:
SAS2-826EL Backplane

 

[shpxnvz]

And since it failed it may be toast anyway as when the import would fail it would delete the pool.

I'm curious about this in particular... is it really expected behavior that if the auto import fails that it deletes data? If so it would probably be good to note that in the instructions, or maybe not recommend auto import in the instructions.

 

[shpxnvz]

Also looking at bug #26834 for that ended volume data loss bug it shows resolved for RC3 which is the version I installed. If I'm reading that right it should no longer be a data loss issue?

 

[gpsguy]

Hold off on doing anything right now.

Let's see if @rs225 can help.

 

[nightshade00013]

Also looking at bug #26834 for that ended volume data loss bug it shows resolved for RC3 which is the version I installed. If I'm reading that right it should no longer be a data loss issue?

Agreed that it should be fixed but even still a fix can always cause a similar bug to rear it's head elsewhere.

But I agree maybe someone who has dealt with the problem before should weigh in on this.

 

[rs225]

I have messaged the OP. Hopefully it will be recoverable, but if this is basically the same problem, that is very troubling.

 

[shpxnvz]

rs225 helped me to get the disk partitions re-added. I ran the commands for all twelve drives ( da0 through da11) and now this is the state of the pool:


root@freenas:~ # zpool import -D
pool: volume1
id: 13367555272724688586
state: DEGRADED (DESTROYED)
status: One or more devices are missing from the system.
action: The pool can be imported despite missing or damaged devices. The
fault tolerance of the pool may be compromised if imported.
see: http://illumos.org/msg/ZFS-8000-2Q
config:

volume1 DEGRADED
raidz2-0 DEGRADED
da0p2 ONLINE
da8p2 ONLINE
da9p2 ONLINE
da1p2 ONLINE
da3p2 ONLINE
da10p2 ONLINE
da4p2 ONLINE
da5p2 ONLINE
17934361591917741187 UNAVAIL cannot open
da6p2 ONLINE
da2p2 ONLINE
spares
13266979735070024154
11910310902942041644


I don't know why da7 and da11 are not showing up. Also don't know if it's safe to run the import command from the CLI or not. rs225 recommended I post up here to see if anyone else may have an idea how to proceed?

Thanks.

ETA:
With rs225's help da7 and da11 are both showing up now.

Code:

root@freenas:~ # zpool import -D
   pool: volume1
	 id: 13367555272724688586
  state: ONLINE (DESTROYED)
 action: The pool can be imported using its name or numeric identifier.
 config:

   volume1				 ONLINE
	 raidz2-0			  ONLINE
	   da0p2			   ONLINE
	   da8p2			   ONLINE
	   da9p2			   ONLINE
	   da1p2			   ONLINE
	   da3p2			   ONLINE
	   da10p2			  ONLINE
	   da4p2			   ONLINE
	   da5p2			   ONLINE
	   da7p1			   ONLINE
	   da6p2			   ONLINE
	   da2p2			   ONLINE
   spares
	 13266979735070024154
	 da11p1

 

[rs225]

My concern was whether 11.1-RC3 might have made any changes to the Corral data in the pool, and what might happen with a failed attempt to bring it in. And what version should be used now?

 

[Ericloewe]

Okay, some thoughts on this:

Importing the pool via the CLI should be 100% safe, since it doesn't touch any middleware code. That allows you to backup the data and do stuff with it, though FreeNAS functionality will probably be impaired.

After the situation is resolved, you might want to wipe the partition table on every drive and resilver it (one at a time, obviously!), to restore the expected format using GPTIDs.

My concern was whether 11.1-RC3 might have made any changes to the Corral data in the pool

Corral is not an issue here, as far as ZFS is concerned. The aclmode thing will not cause any issues at the import stage and is easily corrected.

And what version should be used now?

Good question. At this point, I'd advise against updates in general until this problem has been re-examined. If something happens and you must reinstall, do not import your pool with the GUI before ensuring your backups are in order - you can do so by importing in the CLI and working from there.

By the way, bug ticket for this is here: https://redmine.ixsystems.com/issues/27128

 

[shpxnvz]

Thanks for the reply! Sounds like my best option for the moment is to import from the CLI and transfer data off directly. I'll do some poking around and see if I can import it read-only to minimize the chance of any data getting changed.

 

[Ericloewe]

That's doable, but not needed.

 

[rs225]

I would guess the first import can not be readonly, because the pool is destroyed and needs to be modified back to non-destroyed.

And as @Ericloewe says just now, the risk to data doesn't appear to be there.

 

[shpxnvz]

Just as a follow up, I was able to import the pool without error and am currently in the process of transferring the data off. Thanks everyone for the help.