SCALE is unable to boot or import from GUI when "zpool import" gives warnings

Description

I posted about this issue in the forum .

This is really the combination of a minor bug and a major bug. The minor bug is that replication doesn't copy mount points, so it can create a situation where setting "Read Only" to "Set" in the replication GUI can cause warnings on ZFS import. "zpool import Pool" will warn that it can't create mount points because the dataset is read-only. I've seen this before with TrueNAS CORE, but never experienced any serious consequences from it until now. I also didn't even remember tweaking that replication task, since it didn't cause warnings or issues until I rebooted weeks later.

So, having the above warning when the ZFS pool is imported prevents the GUI from importing that pool. It is fully accessible from the shell despite the GUI saying the import failed, but it doesn't show up under "Storage". What's worse, this also happens during boot, and causes the boot process to stop on "Import ZFS pools" without showing any sort of error message to indicate why it's stuck at that step. In my case the offending pool didn't even have the system dataset, but TrueNAS kept waiting on it to import indefinitely.

Problem/Justification

None

Impact

None

Activity

Show:

Ryan Moeller December 2, 2021 at 6:34 PM

Closing for lack of actionable information.

Ryan Moeller November 11, 2021 at 4:48 PM

Is this fixed on RC 1.1 for people? I'm wondering if this was actually https://jira.ixsystems.com/browse/NAS-113052

Claus Eriksen September 23, 2021 at 8:13 PM
Edited

Just tried disconnecting my pool to see id that would make it possible to pass the import step, but no luck.

Next step will be to test with a fresh install in a new BE and import as had success with.

My system was sidegraded from CORE, and even when I committed to SCALE with a pool upgrade this fails.

 

 

EDIT:

No luck with fresh BE, or even completely fresh install of latest NIGHTLY - going back to 21.08 for now

Rick Bollar September 20, 2021 at 5:42 AM
Edited

I have this issue as well – it occurred when I did a reboot while a pool was dirty (replacing a drive). This pool was created in SCALE, so it doesn't have any legacy feature flag issues.

I worked around the issue by doing a clean install of TrueNAS-SCALE-21.08-BETA.1 into a new boot environment. This successfully started to GUI as you would expect without the pools. I then uploaded a recent backup configuration and the system rebooted back to GUI successfully with the pools already imported. It's interesting that just an older config file allowed the boot to be successful.

The faulty install is in a different boot environment. I assume that once the pool is done resilvering that I could restart back into that environment and capture logs? If so, what would be useful?

Claus Eriksen September 16, 2021 at 5:17 PM

i'm also experiencing this - just updated from 21.08 to NIGHTLY 0916.

Rolled back and things are great again

Need additional information

Details

Assignee

Reporter

Labels

Impact

Fix versions

Affects versions

Priority

More fields

Katalon Platform

Created July 20, 2021 at 12:08 PM
Updated July 6, 2022 at 9:02 PM
Resolved December 2, 2021 at 6:34 PM