You might try setting `sysctl vfs.zfs.recover=1` and `sysctl vfs.zfs.spa.load_verify_metadata=0`.
I had a similar error the other day (couple months ago). The best I did was being able to import the pool read only. I ended up restoring from backup. On Thu, Feb 24, 2022 at 7:30 AM Alexander Motin <m...@freebsd.org> wrote: > On 24.02.2022 10:57, Larry Rosenman wrote: > > On 02/23/2022 9:27 pm, Larry Rosenman wrote: > >> It crashes just after root mount (this is the boot pool and only pool > >> on the system), > >> seeL > >> https://www.lerctr.org/~ler/14-BOOT-Crash.png > > > > Where do I go from here? > > I see 2 ways: 1) Since it is only an assertion and 13 is working (so > far), you may just build 14 kernel without INVARIANTS option and later > recreate the pool when you have time. 2) You may treat it as metadata > corruption: import pool read-only and evacuate the data. If you have > recent enough snapshots you may be able to easily replicate the pool > with all the settings to some other disk. ZIL is not replicated, so > corruptions there should not be a problem. If there are no snapshots, > then either copy on file level, or you may be able to create snapshot > for replication in 13 (on 14 without INVARIANTS), importing pool > read-write. > > -- > Alexander Motin > >
