Bricked (and recovered) 4200
-
Not sure what to say here. When I found that the 25.07 RC has dropped and was a week old I decided to go ahead and take the plunge. My 4200 is essentially new (two months old) and was running the factory image of 24.11 with config restored from the previous sg-5100. I did what I knew (and reviewed in the documentation) to be the proper steps. Running precheck, saving a fresh config, removing installed packages, performing a pre-upgrade reboot. All seemed well -- then I selected the RC and hit upgrade. Everything there also seemed to go well. At conclusion I didn't see any errors, etc.. It rebooted -- and -- nothing. Grabbed my laptop, plugged into the console and it appeared stuck on PXE boot. I hit esc twice to drop to console and tried booting. It would boot a kernel, but then fall into recovery claiming to be missing modules of some kind. It also didn't seem to know anything about an alternate boot image of the previous install. Unfortunately, I didn't have time to really poke at this, so I grabbed an installer from support and reinstalled back to 24.11 with config. All is well save 2 hours of annoyance. This is a home router, but I WFH mainly and between that and my family I have to be careful on downtimes. I can schedule some time and try this again and capture logging. Mainly I just wanted to sound a warning since this is such a new box with a factory install of 24.11. Let me know how you would like me to proceed. Thanks, -JD
-
If it was ZFS it should have been able to roll back to the previous boot environment.
Do you have any logs of the boot failure?
I do know there was an issue with the backend repos yesterday so you may have just hit that. If you can try upgrading again. If it still fails try to grab any logs or console output showing errors.
-
Sounds good. In reality I did something stupid by "just upgrading". If it had not been an RC I would have linked up a logged console, reserved some time to debug, etc. All of which I didn't do. So "boo" to me. It's also possible the image issue you mentioned (I saw it on another thread too) affected the upgrade. I'm in extended negotiations (family :-)) for another downtime this evening. Courtesy of the last attempt I have a re-install key with config handy if things go sideways again. I'll have console logging and other tools on-hand to capture issues. Thanks! -JD
-
Sadly, the new attempt went perfectly. It properly saved a boot env copy, every pkg upgraded without error. Restart was smooth (and log captured). All services are back online. It was back online quickly enough that Zabbix didn't even catch the interface transitions. I would have really wished it hadn't gone this well so I'd have something to give you. On the initial attempt whatever happened was fairly drastic. I'm not sure where to go from here -- you might consider trying it on a fresh factory build 4200 just in case there is some difference on the factory load. I'll be happy to upload my config for you if you want to try it with that. I have a LAGG, several VLANs plus IPv6, but nothing exciting. It might have been the download image hiccup referenced. Otherwise, sorry I don't have more to offer. -JD
-
OK it was almost certainly the backend issue then. Just unfortunate timing the first time you tried to upgrade.
-
I would agree. 18 hours in and everything continues to run smoothly. The issue related to image availability I believe is the valid answer and we can close this out as solved. Thanks everyone. -JD
