> > Is the only way out of this to re-build the filesystem, after
> > shuffling stuff around in a most painful way?
> 
> Yes, unfortunably.

You probably wouldn't wantto  trust the code that would need to be written
in order to add inodes.  Any power fail during the operation would
give you a completely stuffed filesystem.

> > A 32K per inode default seems a bit high IMO.
> 
> Did you change your newfs parameters ? 
> By default it's 8k

Actually it is 4 fragments, so would be 4k for a filesystem with
8k blocks and 4k fragments.

Of course, if you build a large raid set then mkfs could have tried
to give it over 2^32 inodes.  I don't know the effect of this!
(I'm about to commit a patch to fix it.)

	David

-- 
David Laight: david@l8s.co.uk