Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 alpha 4/3/85; site ukma.UUCP
Path: utzoo!watmath!clyde!cbosgd!ukma!david
From: david@ukma.UUCP (David Herron, NPR Lover)
Newsgroups: net.micro.att,net.unix-wizards
Subject: Re: Bizzare paranoia of /bin/login on 3b2
Message-ID: <2144@ukma.UUCP>
Date: Fri, 6-Sep-85 15:46:07 EDT
Article-I.D.: ukma.2144
Posted: Fri Sep  6 15:46:07 1985
Date-Received: Sat, 7-Sep-85 13:30:09 EDT
References: <174@chinet.UUCP>
Reply-To: david@ukma.UUCP (David Herron, NPR Lover)
Organization: Univ. of KY Mathematical Sciences
Lines: 39
Xref: watmath net.micro.att:527 net.unix-wizards:14746

In article <174@chinet.UUCP> djc@chinet.UUCP (David J. Carpenter) writes:
>I have experienced a bizzare problem with logging in on the console
>on our 3b2.  Here's what would happen:  Unix: would send  login: , I
>would reply, then Unix would send Login incorrect immediately without
>asking for a password.  I immediately feared that something happened
>to /etc/passwd.  But I went to other terminals and was able to login
>without any problem.  Then I found that the console would allow me to
>login as guest, which didn't have a password, but from guest, su
>replied with Sorry without asking for a password.  When on as guest,
>the ps command reported that my login -sh was not associated with
>a terminal.
>	The only major change I had recently made of significance was
>to change a line in /etc/inittab from a uugetty to a shell script
>that did a few things, then exec'd uugetty.  Doing a ps from another
>terminal revealed that this shell script was associated with console.

We've experienced this EXACT SAME problem on some of our 3B2's.

BUT, we haven't put any kind of shell script on our console but 
are running getty (not uugetty).  In fact, on one of these machines
we aren't running uugetty ANYWHERE.

We traced the problem to an inability for /dev/tty to be opened
if you're logged in on the console.  getpass() usually opens
/dev/tty to get the password, so that routine will fail.
(Though, doesn't the 4.2 getpass() try other things if /dev/tty
fails?  I don't recall too well).

Unfortunately we can't fire up the machine again because it's sitting
in the other room with no software, waiting for ATTIS to bring us
a fresh core pack.  (There's another problem in the machine where it's
trashing core packs ... anybody else have that problem?)
-- 
--- David Herron
--- ARPA-> ukma!david@ANL-MCS.ARPA
--- UUCP-> {ucbvax,unmvax,boulder,oddjob}!anlams!ukma!david
---        {ihnp4,decvax,ucbvax}!cbosgd!ukma!david

Hackin's in me blood.  My mother was known as Miss Hacker before she married!