Re: Repeated ffs_newvnode panic (dup alloc) on one inode

[Anthony Mallet <anthony.mallet%laas.fr@localhost>]

On Wednesday 10 Dec 2025, at 13:36, mlelstv%serpens.de@localhost wrote:
> The mode is (octal) 142 != 0, so there is some garbage....

Yes, and 0142 doesn't make much sense as a mode. That's also
remarkable that 0142 is consistent accross the different panics.

But doesn't fsdb read the inode from disk? Why doesn't it see the mode
0142? (except if for some reason that was cleared in the meantime but
I don't see how it would auto-clean especially since I had like ~10
panics in a row).


> fsck should be able to clean the inode, you may need to force it when
> it thinks the filesytem was clean.

fsck didn't report any error even though the system still panic'ed
with the same inode afterwards. I also rebuilt the raidframe parity.

> So should the ancient tool clri. It's still a good idea to run fsck
> afterwards.

Is the clri command from fsdb the same as /sbin/clri?