you might have faulty memory. On my uVAX-II this was the case, as you 
describe a problem identical to what i experienced.  Changing a memory
board immediatley fixed the problem. From my reconstruction of what had 
happened (I run Ultrix and 4.3BSD on that machine, so no ddb there :-) 
it was when i tried to do things to that disk that a certain bad piece of
memory got touched, which then set the stage for the panic. 

hope this helps, 
happy hacking
isildur

On Fri, 30 Jul 1999, paul yaskowski wrote:

> I newfs'd a couple of times, and when it finally worked, I mounted the
> drive under /mnt and tried a write test.
> 
> vax# dd if=/dev/zero of=/mnt/test
> /mnt: write failed, file system is full
> Jun 23 19:29:23 vax /netbsd: uid 0 on /mnt: file system full
> dd: /mnt/test: Jun 23 19:29:23 vax /netbsd: uid 0 on /mnt: file system
> full
> No space left on device
> 394545+0 records in
> 394544+0 records out
> 202006528 bytes transferred in 4067 secs (49669 bytes/sec)
> 
> That turned out ok.  Figuring my problem just simply went away, I tried
> installing base.tgz again..
> 
> vax# cat base.tgz | gunzip | (cd /mnt; tar --unlink -xpvf -)
> ...
> ./usr/share/doc/usd/25.tbl
> ./usr/share/doc/usd/26.refer
> ./usr/share/doc/usd/27.invert
> ./usr/share/doc/usd/28.bib
> /mnt: bad dir ino 32258 at offset 512: mangled entry
> panic: bad dir
> syncing disks... 68 67 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 giving up
> ?06 HLT INST
>     PC = 800C18EA
> >>> 
> 
> :(  It's still there.
> 
> > 
> > -- Ragge
> 
> -- 
>  paul yaskowski [a paradigm of a paramount failure]
>