I tried twice to apply the 9.10.2 upgrade (from 9.10.1-U4). The first time, I tried it from a remote location. My internet blipped out, and the web UI lost the connection to the back end. I rebooted. It seemed as if my plugins and jails all booted up properly and seemed to be working, but the webui wasn't there. I ssh'd in and got an error about a missing library (I'm sorry I didn't write it down) when I attempted to sudo. So I ssh'd in as root, and did
Code:
# beadm activate 9.10.1-U4
I rebooted, only to find that I was still in the 9.10.2 BE. Lather, rinse, repeat; same result. `beadmn list` showed that I should have been rebooting back into 9.10.1-U4, it just wasn't happening. This was going to be fun ;-)
I ran the activate command again, this time paying attention to it. It spit out an error stating that grub couldn't be updated. That would seem to explain it. So I did
Code:
#mkdir /tmp/old-be
#beadm mount 9.10.1-U4 /tmp/old-be
figuring that I could grab the grub.cfg from the old BE, drop it in place, and reboot again. No such luck, since grub.cfg didn't seem to be in the expected place.
I finally got out of the bad BE with the following:
Code:
# cp /boot/grub/grub.cfg /boot/grub/grub.cfg.bak
# cat /boot/grub/grub/cfg | sed 's/9.10.2/9.10.1-U4/g' > /boot/grub/grub.new
# mv /boot/grub/grub.new /boot/grub/grub.cfg
# chmod 600 /boot/grub/grub.cfg
After a reboot, I was back in 9.10.1-U4, and I deleted the 9.10.2 boot environment.
That was yesterday. Since I assumed that all of this breakage was my fault for attempting an upgrade over a VPN, I decided to try again this morning, when I am on the same network as the FreeNAS box.
The upgrade job is once again stuck at 20% (installing base OS), and has been sitting there for about 30 minutes now. When I ssh in, I see the following:
Code:
#beadm list
BE Active Mountpoint Space Created Nickname
default - - 931.3M 2015-01-31 03:34 default
9.10.1 - - 884.4M 2016-08-10 18:30 9.10.1
9.10.1-U2 - - 1.3G 2016-10-11 07:12 9.10.1-U2
9.10.1-U4 NR / 6.3G 2016-11-10 10:04 9.10.1-U4
9.10.2 - /var/tmp/tmp9NF_Gz 13.2M 2016-12-23 07:27 9.10.2
I don't see any sort of failed update log in /data, nor do I see anything particularly alarming in /var/log/messages. There was some stuff that is probably related in /var/log/debug.log. I put what I think is relevant in a pastebin:
http://pastebin.com/1KrQTR5p
As you can see, if this update succeeded, it didn't correctly set the boot environment, nor did it automatically reboot. I'll keep this BE around in case there is anything else I can do to help assist with this. I'm also in the IRC channel on Freenode.
UPDATE: Several hours later ....
I left the thing, and took off to run some errands. I came back, and it looks as if the update completed. Based on the current uptime, it looks as if the update took around 2.5 hours to apply itself. That's a long time.