On Fri 30 Mar 2001 at 16:09:41 -0800, Chuck Silvers wrote:
> hi,
> 
> is this repeatable?  

It is repeatable, it seems. I again copied the same NetBSD source tree
over the partial copy that I already had, and this time I think it
paniced sooner. Now it stopped approximately in
regress/lib/libcrypto/bf, while the first time it also got through the
whole of sbin, usr.bin, usr.sbin, gnu, half of share, and missing sys
altogether.

ddb has no working stack backtrace on the VAX, unfortunately. When I
tried to sync, I got a new panic:

db> sync
syncing disks... Trap: type 2, code 0, pc 80000430, psl 4080000
panic: trap
Stopped in tar at       _arithflt+0x131:        pushab   gcc2_compiled.+0x111e
db>

An attempt to dump (even if the result would be useless) gave, much
too quickly:

db> call cpu_reboot(0x100)

dumping to dev 20,1 offset 197144
dump succeeded

and the machine halted. I don't think it actually did a dump.

After reboot, fsck had a lot of bad block counts and unreferenced files
to clean up. Savecore detected no dump.

I used the following script to copy the tree, from an nfs mounted
source:

#!/bin/sh
Usage="Usage: cptree fromdir todir"
from=${1:?$Usage}
to=${2:?$Usage}

[ -d $from ] || { echo "$from is not a directory."; exit 1; }
[ -d $to   ] || { echo "$to is not a directory."; exit 1; }

(cd $from && tar --one-file-system --sparse -cf - .) |
        (cd $to && tar --unlink --sparse -xvpf -)

> -Chuck
-Olaf.
-- 
___ Olaf 'Rhialto' Seibert - rhialto@polder --Soep van de dag, wat zal dat zijn
\X/ land.nl     --wat kan dat wezen, beter maar het ergste vrezen -Boy Bensdorp