Check your pools. This update hosed my system also. It refused to mount the pool with pretty much everything including the system dataset in it.
I figured it out though. My system consists of 5 14tb drives in raidz1 configuration (so I lose one to parity). I also have 2 additional drives plugged in via USB for the read and write cache. They are nvme drives in enclosures. I added both of them to my primary pool to help with caching.
Upon reboot, the pool wouldn't mount. I nearly panicked but this is a newly built system, running reliably until it hit the Python bug last night and core dumped, killing the UI but continuing other services. I was happy to see a point update that appeared to address this but it blasted my pool offline.
When I issued the command "zfs mount", it gave me a list of available pools, showing my main pool twice; because the cache drives serve the pool, it saw two identically named pools. Looked like an overlooked bug to me. Unplugged the write cache and rebooted. Same problem, pool didn't mount. Unplugged the other one, zfs mounted the pool, voila. Just to fully test it, I rebooted again, and the pool mounted again, but was degraded until I kicked the last cache drive out of the pool. Now it's happy and it'll survive reboots and bring the pool online as expected.
In 7 years of running Freenas/Truenas, this is the first show stopping upgrade I have ever seen.
