panic: ffs_blkfree_snap: bad size (was: /dev/fss0: file system not clean)

For some time now, I'm getting (10.1/amd64)
> /dev/fss0: file system not clean (fs_clean=0x2); please fsck(8)
when taking a snapshot and mounting that (for making a backup).

Today, I additionally got (shortly after that message for the /var snapshot)
panic: ffs_blkfree_snap: bad size: dev = 0x1205, bno = 8842446292672000718 bsize = 32768, size = 32768, fs = /var

I rebooted and got
/var: replaying log to disk
[...]
panic: ffs_blkfree: bad size: dev = 0x1205, bno = 377710485266874556 bsize = 32768, size = 32768, fs = /var

I rebooted single user and fsck'd /var
8842446292672000718 BAD I=16444357
-6938471369708809904 BAD I=16444357
[...]
INCORRECT BLOCK COUNT I=16444357 (300416 should be 241536)
[...]
BAD/DUP FILE I=16444357 OWNER=0 MODE=100600
SIZE=536870912008 MTIME=Aug 22 09:17 2025
[...]
CANNOT READ: BLK 205532999823616

THE FOLLOWING DISK SECTORS COULD NOT BE READ: 205532999823616, 205532999823617, 205532999823618, 205532999823619, 205532999823620, 205532999823621, 20553299982:
622, 205532999823623, 205532999823624, 205532999823625, 205532999823626, 205532!
99823627, 205532999823628, 205532999823629, 205532999823630, 205532999823631, 21
5532999823632, 205532999823633, 205532999823634, 205532999823635, 2055329998236:
6, 205532999823637, 205532999823638, 205532999823639, 205532999823640, 20553299!
823641, 205532999823642, 205532999823643, 205532999823644, 205532999823645, 205!
32999823646, 205532999823647, 205532999823648, 205532999823649, 205532999823650.
205532999823651, 205532999823652, 205532999823653, 205532999823654, 20553299982
3655, 205532999823656, 205532999823657, 205532999823658, 205532999823659, 205536
999823660, 205532999823661, 205532999823662, 205532999823663, 205532999823664, і
05532999823665, 205532999823666, 205532999823667, 205532999823668, 2055329998236
69, 205532999823670, 205532999823671, 205532999823672, 205532999823673, 2055329!
9823674, 205532999823675, 205532999823676, 205532999823677, 205532999823678, 201
532999823679,
CG 135247552: ALLOCBLK: BAD MAGIC NUMBER
WRITING ZERO'ED BLOCK 205532999823616 TO DISK
CANNOT WRITE: BLK 205532999823616
[...]

I gave up and fsck'd -y /var, which gave the same messages plus several 
screenfulls of by-scrolling messages like the last one and then
FREE BLK COUNT(S) WRONG IN SUPERBLK
SALVAGE? yes
SUMMARY INFORMATION BAD
SALVAGE? yes
BLK(S) MISSING IN BIT MAPS
SALVAGE? yes

I fsck'd -f again, getting the Phase 5 messages again (and slightly different 
summary information at the end). Another fsck -f didn't find anything more.

I then went multi user and it seems to work.

What the hell is that? How can I debug it?