UI and GUI crash after Removal of 3 disks on ZFS2

[Hellrazorx]

Hi,

I've been testing a fresh install on a rig on which I setup
-A pool
-A dataset
-A smb Share (I didn't test while this one is disabled tho)

I can freely disconnect two drives at a time with no problems at all, Swapping and all.

If I ever disconnect a third drive, the pool gets offline. (expected)
But the web GUI goes ''System disconnect or blabla''
And I can get the SSH prompt but freezes when trying to input password.

On the physical server, I can ask for a reboot but will take forever to close pending tasks.

***

I can reconnect all drives and do a hard reset physically. Everything goes back to normal after that.
But I think this is not official procedure!
***
Note: Just to point out, I am aware that the possibility of three drives dying in such a short span is nearly impossible.
I'm mostly thinking this may cause a problem if I use multiple pools, if for some reason a small ZFS array dies, why would the entire system be compromised?



Thanks a lot for your inputs.

 

[sretalla]

Where is your system dataset?

 

[Hellrazorx]

Where is your system dataset?

Onto two separate SSD's, selected at install

If I manually reboot the system when it's feezed (Hard reset, no choice), I can get back access to the UI once it rebooted.

I also tried doing the test when disabling SMB services, the system crashes the same way.

 

[ginjiruu]

Having a similar issue doing some initial tests. However it happens whenever a vdev goes completely offline. eg we removed both de-dupe drives from our system causing the webgui to become inaccessible. Any ideas why this happens?
Rebooting also fixed the webgui in our case

 

[Patrick M. Hausen]

Onto two separate SSD's, selected at install

Not the boot pool. The system dataset. Please check in the settings: System > System Dataset.

If it's on the storage pool you forcefully offlined, the system comes to a halt. This is expected. If you run your boot pool on a pair of SSDs you can put the system dataset there.

 

[danb35]

Also, if you were using swap at the time, pulling drives could cause problems there. Swap is mirrored, but loss of more than one disk greatly increases the chances of a mirror failing, and thus swap failing.

 

[ginjiruu]

Not the boot pool. The system dataset. Please check in the settings: System > System Dataset.

If it's on the storage pool you forcefully offlined, the system comes to a halt. This is expected. If you run your boot pool on a pair of SSDs you can put the system dataset there.

Changing the system dataset fixed it for me. Thanks a ton!