this is way slower than it should be. ssh on a microvax will be slow,
but not THAT slow.
if its the first tim ever connecting, generating new keys will take some
time.. but on my KA650 (NetBSD 1.3.3, 16MB core, RF72 disk on a KFQSA 
(which is itself the epitome of slowness!)) it took about 3 or 4 
minutes to generate the key, and after that, ssh takes about 20-30 sec 
and then its decent speed once connected. you can tell if you do something
that dumps a lot of output, that ssh is noticeably slower than a telnet, 
but it is still quite usable. 
youre using it as a console server? nice, thats exactly what i'm using
the machine i described for.. one CXY08 in it right now, and a second
going into it soon.. to serve (and log) the consoles of half a dozen 
alphas and 4 or 5 sparcs. :-)
one final comment which just occurred to me now: the ssh i am running on 
that machine is _not_ the openssh from the netbsd packages.. it is an
older ssh i compiled myself.

happy hacking,
isildur

On Tue, 11 Feb 2003, Charles H. Dickman wrote:

> I have a KA650 processor in a BA23 chassis (MVIII?), that I want to use 
> as a console server, because with a couple DHV11's, it has LOTS of 
> serial ports. I am running a 1.5 kernel and logging in with ssh is VERY 
> slow. From the remote ssh login, to a request for a password is close to 
> 10 minutes. During the entire time, the load on the uVAX is 100%. After 
> that, the ssh shell response is not noticably slow. Any suggestions? Or 
> is this what I must expect from her?
>