Force-degrade pool

[vitaprimo]

Since my ZFS pool fails to mount, I thought maybe I could nudge it into working by changing enough things about it so the system is forced to look into it…or something. It's a ZRAID, it should be able to work with any 1 missing disk. I removed 1 disk and booted the system.

 

I log in to mount it manually the pool (auto-mount is temporarily disabled) but it tells me I have a missing cache disk and won't let me start the array without extra steps. I'm not sure what it's talking about since I don't have a cache anything; it's an all-flash ZFS pool, there's no need. I have one other obligatory disk in the regular array but it's empty and doesn't have any of the supporting pools (parity, cache).

 

 

 

Does it refer to this pool – the ZRAID one – as the cache?

 

And followup: when it says "remove the missing cache disk" does it mean the individual disk drive or the volume "disk" would be?

 

And followup of the followup: if it means the whole pool, how do I get it to mount degraded then?

 

I can't remove one disk while mounted to degrade it live because that's why I'm trying to get it to mount degraded, it won't mount when it's normal.

 

C'est tout, thanks.

[JorgeB]

1 hour ago, vitaprimo said:

ZRAID

I assume you mean raidz?

 

Please post the diagnostics and the output of:

zpool import

 

[vitaprimo]

Oh yeah, my bad. I just caught on. Thanks for correcting me.

 

I'll do that and come back. It might take a little bit, I had reinserted the ESXi SD card to check on something. It goes inside the giant heavy toaster. 😔

 

Thanks!

[vitaprimo]

It shows up !

 

Last login: Wed Nov 22 03:15:25 on ttys003
[Wed22@ 5:36:29][v@zx9:~] $〉ssh zx3
Linux 6.1.49-Unraid.
[Wed22@ 5:36:32][root@zx3:~] #〉zpool import
   pool: disk1
     id: 9807385397724693529
  state: ONLINE
 action: The pool can be imported using its name or numeric identifier.
 config:

	disk1       ONLINE
	  sdd1      ONLINE

   pool: alpha
     id: 1551723972850019203
  state: DEGRADED
status: One or more devices contains corrupted data.
 action: The pool can be imported despite missing or damaged devices.  The
	fault tolerance of the pool may be compromised if imported.
   see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-4J
 config:

	alpha       DEGRADED
	  raidz1-0  DEGRADED
	    sdb1    ONLINE
	    sdc1    ONLINE
	    sdd1    UNAVAIL  invalid label
[Wed22@ 5:36:35][root@zx3:~] #〉zpool list
no pools available
[Wed22@ 5:38:17][root@zx3:~] #〉

 

sdd1 is not inserted, BTW. No idea why it says invalid label. Now I just need to figure out the order to start, if I should start at all, the array.

 

I think it should be mountable without starting the array with zpool import alpha though I'll continue reading — I'm using the manpages from Fedora 39's ZFS — a bit more for clues, then I'll skim Unraid's docu one last time. It's very little VM data, of which I have a backup, or rather a version, but the one in these disks has been OCPDed to the max.

 

Thank you ! ❤️

[vitaprimo]

Scratch that, I couldn't wait.

 

I did zpool import disk1 to import the main array without starting the thing, it succeeded. Therefore, that's the asnwer to the question.

 

However, in regards to what I wanted to do, it didn't quite work, the problem that appeared earlier when attempting mounting the pool the system would hang forever (and show some micro kpanics) is still there.

 

[Wed22@ 6:15:58][root@zx3:~] #〉zpool import alpha

Message from syslogd@zx3 at Nov 22 06:16:46 ...
 kernel:VERIFY3(size <= rt->rt_space) failed (281442912784384 <= 2054406144)

Message from syslogd@zx3 at Nov 22 06:16:46 ...
 kernel:PANIC at range_tree.c:436:range_tree_remove_impl()

 

I haven't given up though, I think I still might have an idea or two.

 

Thanks! =]

[JorgeB]

Pool looks corrupt, try mounting it read-only:

 

zpool import -o readonly=on pool_name

 

[vitaprimo]

Thanks,

 

I'm sorry, I meant to answer earlier but I got food poisoning, slept-on-the-shower-food-poisoning. 🤢 On the plus side, I may have lost some weight.

 

I'm not using the disks anymore, even if they have not shown any signs of failure, I know they old, but pretending they're good I kept trying different things so I'd know how to proceed when an actual emergency presents itself; the disks wouldn't mount again formatted as ZFS either in their own pool or in the Unraid pool, as a group or by each of themselves (after Tools→New Config, of course).

 

I remembered though, that ZFS format data can persist on a disk even after this has been reformatted to something else, so I changed the format to Btrfs and finally they mounted again. I don't know what to make of it, I'm just leaving it out there for whomever it serves a purpose.

 

Gratzie ancora. 🙇‍♂️