Our Latest Posts: -  MacScan  -  Disaster Averted  -  Network Troubleshooting  -  Bad Attitude

Translators wanted!

---

Best of the Newsgroups:

---

See http://aplawrence.com/Bofcusm/2648.html for what this actually was..

What is this stuff?

Main Index

From: Bela Lubkin <f...@armory.com>
Subject: Re: Users Can't Login on the Console (tty01)
Date: Sun, 12 Nov 2006 15:22:43 -0800
Message-ID: <200611121522.aa01474@deepthought.armory.com> 
References: <20061112200651.GA22075@trebor.iglou.com> 

Bob Stockler wrote:

>  System:  SCO OSR6 (MP1, OSS702a, OSS702b and MP2)
> 
> Problem:  No user can log in on the console (tty01).
> 
> This just started recently, after a reboot following a kernel
> panic.  The Software Manager (custom) has no complaints after
> verifying all software installed on the system.
> 
> All users can log in on tty02-tty12 and, root has no problem
> logging into single user mode, but in multiuser mode . . .
> 
> At the initial "login:" prompt I must type each character of
> the user name and the [Enter] key twice before the "Password:"
> prompt appears.
> 
> After typing in the password I must press the [Enter] key two
> times.  I doesn't matter if I single or double strike the  
> characters in the password, I always get:
> 
>   Login incorrect
> 
>   Wait for login retry .
>   login: _
> 
> At this point it accepts single striking the characters in the
> user name, but the "Password:" prompt never appears.  I finally
> get:
> 
>   Login timed out
> 
>   Wait for login exit:
> 
> after which the screen clears, the initial "login:" prompt
> reappears, and the cycle repeats itself.
> 
> If I press [Enter] before seeing the "timed out" error message
> I simply get another "login:" prompt.  Entering the user name
> at it evokes the same behavior as did the previous instance.

That description sounds just like two processes are competing to consume
your input.  Does OSR6 come with `lsof`?  Login on a tty that does work
and run `lsof /dev/tty01`, what processes have it open?  Compare to
output for another not-logged-in console tty that you know isn't having
the problem.

It's common for multiple device nodes to exist which access the same
tty.  Again, I'm not that familiar with OSR6, but see if any other names
also mean "tty01".  e.g. /dev/console, syscon, systty, vt00 or vt01,
etc.  You can test by sitting at a root shell prompt on tty02 and doing:

  echo checking syscon > /dev/syscon

etc. (only do this to names that already exist in /dev and that you have
at least some plausible reason to believe might be aliases for tty01).
Anyone, once you've done a series of those, then flip the display to
tty01 and see which ones showed up.

Now that you have a list of aliases for tty01, run `lsof` on them:

  lsof /dev/tty01 /dev/console /dev/vt00 ...

You'll see at least one other process besides the expected `getty`.  Now
figure out why it is running -- wrong entry in inittab?  rc script gone
wild?

Once you've identified it, see what happens if you kill it.  Does
something (probably `init`) start another copy?  Then you know where to
look (inittab).  Does it die and stay dead?  Look for an rc script or
other non-repeating instruction.

>Bela<