Update didn't reboot.. did it finish?

[j_s]

So I just tried to update from 22.7.11_1 to 23.1(.x???) via the WebGUI.  From the WebGUI everything looked pretty normal, except the reboot never happened.

From the WebGUI:

***GOT REQUEST TO UPGRADE***
Currently running OPNsense 22.7.11_1 (amd64/OpenSSL) at Tue Feb  7 00:17:51 MST 2023
Fetching packages-23.1-OpenSSL-amd64.tar: ......................................................... done
Fetching base-23.1-amd64.txz: ............. done
Fetching kernel-23.1-amd64.txz: ...... done
!!!!!!!!!!!! ATTENTION !!!!!!!!!!!!!!!
! A critical upgrade is in progress. !
! Please do not turn off the system. !
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
Extracting packages-23.1-OpenSSL-amd64.tar... done
Extracting base-23.1-amd64.txz... done
Installing kernel-23.1-amd64.txz... done
Please reboot.
***REBOOT***
***DONE***

Okay, so it's been more than 20 minutes and the system hasn't rebooted.  Uptime is 12+ days.

IIRC when I upgraded from 22.1 to 22.7 I did a reboot when i thought something was wrong with the upgrade, and the system was rendered unbootable.  Now I'm wondering what I should do next, and if I should gather some logs or anything before going further.  At the time I blamed myself for being too impatient, but this time I'm much more patient and trying to avoid having to do a reinstall and upload the config file again.

Note that during the entire 22.7 lifetime, if an update required a reboot, it always went off without a hitch.

So what should I check or do as next steps?  I feel like the next step is to reboot, but if I'm going to reboot and have an unusable system again, I'd really like to know what is going on.  Any log files that might be useful for troubleshooting?

[j_s]

So about 2 minutes after I posted this the system rebooted.  It then took over an hour to do the reboots and install the OS.  Upgrade on first reboot was slow.  Bootups were slow.  Package installs and updates were slow.  I waited it out (very patiently) and it did eventually come up and "just work".

Digging deeper, once opnsense was booted back up and fully functional, my SSD write speeds were around 1MB/sec maximum (nope, that's not a typo).  After some more thinking and analyzing, ZFS property autotrim was set to off.  As I was never running the "zpool trim <poolname>" command from the command line, I assume it needed a trim.  After a zpool trim and waiting a few minutes, I could write at over 100MB/sec.

I've since changed the autotrim property to on with the command "zpool autotrim=on zroot".  Probably would be a good idea for the default to have autotrim set to on.  As a ZFS guru, I will admit that there are situations where autotrim=off may have it's place for some situations and some workloads, I suspect the greater masses would benefit from autotrim being set to on when a zpool is created.

Edit: For anyone that is curious, the total upgrade time from when I first clicked "update" in the WebGUI to the system booted up and functional was almost 2 hours.  LOL.

[cayenne]

same problem since a few months, forced to launch the reboot in console