Hi all,
I have a strange issue with TrueNAS Scale 25.10.0.1:
When I try to update to 25.10.1 it goes through the motions, tells me to restart, creates a new boot environment and restarts. But when it’s up again, the version still shows 25.10.0.1 and an available update and the new boot environment is set to be default on restart but the 25.10.0.1 is currently active.
I tried it many times, deleted the one app that was on the system, tried the manual update and restarted end even shut down many times. This is the only system with issues and I can’t recall having those issues ever and I’m a user since the FreeNAS days.
I’d appreciate any help. I don’t really want to reinstall the system as a whole. Does anybody have any ideas what the issue could be?
Have you deleted the 25.10.1 boot environment while you were running 25.10.0.1 and tried the update process again?
If you download your current system configuration, do a fresh install of 25.10.1 and reload your configuration you should be back to normal. It’s not like you have to set up the whole system again.
If you have a spare boot device, you could power down, swap boot devices and do a clean install and system configuration reload. You would have the previous boot device to swap back if you didn’t like it.
You can try running scrubs on your boot-pool and data before upgrading, if you haven’t recently.
Thanks for you suggestions and quick reply.
Yes, I tried to delete the boot 10.1 environment but it didn’t change the behaviour.
As for your other suggestion; I know, but thanks anyway.
I really wanted to solve that issue since it bugs me and I want to find out what the issue is. Also, I can’t really access the system until the weekend again. But I know that just reinstalling and restoring the config might be the fastest way all together.
The boot-pool is scrubbed every week automatically, so that unfortunately isn’t a solution.
I’ll investigate further and will update here if I find a reason or solution.
Search the forum for your issue. It may be one that you wait for 25.10.2 for the update. There may be tickets listed in the Jira system also.
I searched the forum, and the internet already but I didn’t think of Jira, thanks for the hint.
I re-searched the forum (wow, what a difference a dash makes) now and Jira. I couldn’t find anything related to my issue. There was one old forum post with a similar issue which ended with no further answers and a stumped DjP-iX being left unanswered. There was one with different devices not seeing the grub entries which could in theory under very weird circumstances be related to my issue, but this I can only investigate when I can access the system again.
Regarding Jira: admittedly, I didn’t open all tickets containing the word update, but I look over all subjects and opened a few that seemed to make sense. Still, nothing that helped me.
Did you just mention those things because most users never bother to look or did you actually find something? If so, could you point me in the right direction, since I really tried for some days now and was unable to find anything?
There were a few mentions of update trouble. Two were problems with running out of space for UEFI while the boot-pool had enough space. There have been others with update trouble. First fix is deleting the bad boot environment and trying the update again. Next is trying to do the update with a manual update download. Last was usually the download of a fresh install with restoring the system configuration.
I mentioned searching because you might find something that matched your pattern. If you have more previous boot environments, you could try booting to those and attempt update
I’ve ran into this issue rarely on nightly updates. I usually extract the TrueNAS-SCALE.update file from the iso and install using the Manual update option. If all else fails, backup your settings, create a USB boot from the iso and install via boot. Either options have always worked for me.
Unfortunately neither booting from old environments nor the manual update helped. I tired those before in various combinations. But I’ll keep an eye on the forum and Jira. Fortunately for everyone else it doesn’t seem to be a common issue or bug.
After reading through the forum posts before, I have a hunch that it might be a failed old boot device already ejected from the pool but still physically attached. When trying to interact with it in the GUI, it is stated that it is a read-only device. IF the updater for whatever reason tries to update something on this device, it would make sense that it doesn’t work. It wouldn’t make sense that the updater tried this useless task. But who knows. I need to read up on blocking devices from the OS. I know how to hide disks, never tried the whole device. Or I wait until the weekend when I can physically mess with the system. We’ll see.
Thanks for your inputs though! Even if I tried all or most of it before, it still forces new perspectives and to think about the issue maybe from another angle.
So, I found the issue/solution:
As expected, the failed boot drive which was still physically attached prevented an update. The reason was, that boot devices are being set to read only (most of the fs anyway) since Scale 24.00 I think. Somehow, TrueNAS could not update the boot environment fully with the failed device which looked like it recovered since the removal but was in a weird state still being a boot device and thus read only while also not being part of the boot pool. After removal of said device the system booted in the updated boot environment as it should have after the update.
Why the boot device recovered, I don’t know. But it was removed in a failed state. When it recovered, nothing on the drive indicated that it had failed, it seems. This is what might have caused it to prevent a proper update. I might file this as a bug in Jira since it seems to affect failed and recovered boot devices since TrueNAS 24.00.
So for anyone having this issue a TL;DR:
- Old boot device might prevent update, since it is read-only, even if it is not part of the boot pool anymore.
- Removing the old device physically after removing it from the pool solves the issue
2 Likes
