> 
> 	NetBSD dump seems to default to 10k (which maps very nicely to
> 	the 20 blocks) per write.

OK, that seems reasonable.

> 	What happens if you use -b to switch the blocksize to say 63
> 	(which might be the SunOS default) or even higher?

I tried that, and it still went one write of blocks, then slept for
30 seconds, then another, then sleep for 30 seconds.

> 	Manuel Bouyer recently commited changes to dump to use a read
> 	cache, which speeded things up quite a bit for normal use. It
> 	might help your situation. Presumably the next write is not ready
> 	by the time the drive has finished with the previous :/

I suspect it might be timing related, but I am not sure, presently.
I did change cpu's last night from a 3/160 to a 3/260, and I will
try that, setup tonight and compare.  It has 32M ram, and maybe that
might make some difference.

Also, I have a spare wangtek that I will put in in place of the
archive, just in case it may be drive hardware related.  I have
seen something that differentiated the two drive types in sun
bits.  Maybe there is something different in the way the hardware
should be accessed.  Dunno, yet.

Thanks
Bob Keys