5810 installation report

[Alan's Home for Wayward Notes File.]

In article <9179 at cbmvax.commodore.com>, grr at cbmvax.commodore.com (George Robbins) writes:
>
> [ Lots of stuff deleted. ]
>
> Second, I've got a couple of DMB32's, but when I activated logins on a
> couple of lines (with no cables attached yet) a later caught them running
> up a bunch of CPU time.  This isn't funny, but I haven't had a chance to
> try various thing and see what the system thinks it's doing.
> 
> Anybody had problems with DMB32's on VAX Ultrix?

	I've seen this sort of problem of nearly every version of
	ULTRIX and combination of CPU and terminal controller since
	the beginning of time.  It goes something like this:

	1.  The getty with the line open, sees noise and thinks
	    it's someone trying to login, so it either prompts
	    for the user name or forks a login to let it do the
	    work.

	2.  Login searchs the password file for the noise and prompts
	    for a password.  Even if it doesn't find a username it 
	    asks for a password so people breaking in don't know if 
	    the account is valid.  Since the username was noise the
	    login is invalid, so login exits.

	3.  Some time later init notices that nothing is running on
	    on the tty line, so it starts another getty.  Goto #1.
	    Init is (or should be) smart enough to see if this happens 
	    with great frequency and scribbles a message to syslog,
	    and waits a short time before restarting the getty.

	4.  Symtoms: Lots of system calls, lots of time spent in
	    kernel mode and lots of forks.

	    Solution: All activeated tty lines should be connected
	    to terminals and the terminals should be turned on.
	
>
> -- 
> George Robbins - now working for,	uucp: {uunet|pyramid|rutgers}!cbmvax!grr
> but no way officially representing	arpa: cbmvax!grr at uunet.uu.net
> Commodore, Engineering Department	fone: 215-431-9255 (only by moonlite)

-- 
Alan Rollow				alan at nabeth.enet.dec.com