TrueNAS SCALE 23.10.2 has been released!

[morganL]

It may be a few days before "time at home" and "can tolerate apps being down for several hours" will coincide, and I have no confidence at all that the system will be able to generate a debug file--but if it can, I'll try to get a ticket filed.

Any chance the bug you saw was tied to the GPU?

Applications will not start after upgrading to 23.10.2

Interesting note - may be it will help someone: after upgrading from 23.10.1 to 23.10.2 I got "Failed to start kubernetes cluster for Applications: server is not initialized yet" and all apps are in "deploying" mode. Waiting for over an hour and reboot did not help. What had solved the issue -...

www.truenas.com

 

[danb35]

Any chance the bug you saw was tied to the GPU?

Applications will not start after upgrading to 23.10.2

Interesting note - may be it will help someone: after upgrading from 23.10.1 to 23.10.2 I got "Failed to start kubernetes cluster for Applications: server is not initialized yet" and all apps are in "deploying" mode. Waiting for over an hour and reboot did not help. What had solved the issue -...

www.truenas.com

Unlikely--the only video in that system is associated with IPMI. There's no separate GPU, no passthrough, etc.

 

[danb35]

Tried the upgrade from 23.10.1 to 23.10.2 again. After booting into .2, the apps screen was showing this error:
(seriously? Screen shots aren't allowed here, editing posts isn't allowed here--that really limits the usefulness of this forum)

Error in Apps Service
Application(s) have failed to start:
[EFAULT] Failed to configure PV/PVCs support: Cannot connect to host 127.0.0.1:6443 ssl:default [Connect call failed ('127.0.0.1', 6443)]

(no apologies for typos; if this forum allowed images they wouldn't be present)

Notably, that error wasn't present when I'd previously tried the upgrade. Then I followed the link you gave and saw that it really didn't have anything to do with having a GPU, but with K3s GPU support. I unchecked that box, saved, and in due course I see "Apps Service Running." But nearly an hour after that, only two apps are showing "running", even though they aren't, really: prometheus-operator from TrueCharts, and a "custom app" of UrBackup. 31 of the other 33 remain in "Deploying"; the other two aren't supposed to start anyway.

Meanwhile, I'm still getting periodic alert emails about failure to connect to the middleware. Ticket created:

Jira

ixsystems.atlassian.net

New is that the middleware is so wedged it can't even give me the list of boot environments. And the console is spammed so hard with networking messages (which is a problem you've had for over a decade, spamming the console should be reserved for issues on the level of "your server is on fire") that I can't even use the console menu to reboot. But I was able to ssh in and reboot, and choose the 23.10.1 boot environment from the GRUB menu. And as before, reverting to 23.10.1 results in the apps all starting again.

 

[Daisuke]

Thank you for another perfectly uneventful release. Server up and running in less than 3 minutes, all TrueCharts apps functional.

 

[thaddeusf]

Build seems to require HBA in IT mode. Discovered I was using IR mode and cannot yet flash for bunch of reasons.
The most troubling issue is when I select 23.10.1.3 to recover from in boot menu, the upgrade broke this as well. Could the zfs driver have updated other system or the dataset? I've been down a week now. Probably would have been a good idea to have this as a major issue. Surely I'm not only one....

 

[danb35]

Probably would have been a good idea to have this as a major issue.

If you think so, it'd be better to log it under "Report a Bug" at the top of the page. With a debug file would be better yet.

Surely I'm not only one....

First one I've seen, though that doesn't prove much.

 

[thaddeusf]

Build seems to require HBA in IT mode. Discovered I was using IR mode and cannot yet flash for bunch of reasons.
The most troubling issue is when I select 23.10.1.3 to recover from in boot menu, the upgrade broke this as well. Could the zfs driver have updated other system or the dataset? I've been down a week now. Probably would have been a good idea to have this as a major issue. Surely I'm not only one....

@here. My assumption seems to be false. There was something in my configuration that caused the problem. I'm doing a fresh install and I'm able to use .20. I'm at a loss how something transferable in the configuration backup could cause the behavior.