> > "ext2fs: Warning: bad directory entry: XX inode: YY offset: ZZ"
> Argh, I just realize that i saw this, too!
> It was a few days ago, when I extracted a hurd package under
> Linux with dpkg-cross. Booting resulted in a hung system, because
> serverboot could not access /hurd/ext2fs.static. efsck under Linux
> did NOT report anything. Linux could access the directory just fine.
> The XX was 119, the inode did NOT exist when asking debugfs.
This is almost identical to the behavior I have been experiencing,
although of a lesser degree.
I also see XX=119, but I have also seen XX=255. I am not sure if
there is any significance to this.
The problems seem isolated to the particular partition in which the
heaviest disk activity is occuring. However, I believe that in
some cases corruption of the root system takes place, because I
have observed corruption of some base libraries rendering certain
executables non-functional. This may only occur when installing
packages, but I can't be certain.