Xterms not cleaning up utmp properly / Telnet woes

Xterms not cleaning up utmp properly / Telnet woes

Post by Nick Hillia » Mon, 22 Feb 1993 07:10:33



Hi,

When I log out of an xterm, it doesn't seem to clean up its entry properly
in /etc/utmp, so that when someone fingers the system, it appears as if I'm
still logged on. Does anyone know how to fix this problem? (I'm using SLS,
kernel 0.99pl4).

BTW, has anyone else noticed that telnet sessions into linux hang a lot
more in 0.99pl5 than in, say, pl4?  NCSA telnet from a PC seems to hang
pretty much at random, which is a Bad Thing(tm) if you're trying to teach
people UNIX.  Telnetting from VMS machines is even worse. Sometimes I can't
even log in before it crashes!  No problems using NCSA, though, with 0.99.4.

I'm using 2 ethernet cards, which could be the problem. A WD8013 and a
D-Link 220 (NE2000 compat) with Don Becker's drivers. As I say, fine with
0.99.4, but no luck with 0.99.5 :-(

Thanks in advance,
Nick
--
| Nick Hilliard            |                                                |



 
 
 

1. Problem with 2.2.6, xterm failing to update utmp properly

I've just installed a new 2.2.6 machine with xfree86.

What I've found is that when closing down an xterm it fails to remove
the utmp entry (at lest that's my guess), so you still get 'w' showing
open pty's:

% w
 3:32PM  up 5 days,  6:28, 21 users, load averages: 0.38, 0.29, 0.21

[...]
paul     p8  :0.0              1:02PM  2:30 -
[...]
paul     pj  :0.0             Mon01PM 26:02 -

(The current command showing as '-' demonstrates what I mean).

Running a ktrace on xterm doesn't show any problem:

  1726 xterm    CALL  open(0x26028,0x1,0xa)
  1726 xterm    NAMI  "/var/run/utmp"
  1726 xterm    RET   open 5
  1726 xterm    CALL  lseek(0x5,0,0x3a8,0,0)
  1726 xterm    RET   lseek 936/0x3a8
  1726 xterm    CALL  write(0x5,0xefbfd6d8,0x24)
  1726 xterm    GIO   fd 5 wrote 36 bytes
   "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"
  1726 xterm    RET   write 36/0x24
  1726 xterm    CALL  close(0x5)

My guess is that something messed up during the installation (I went
through some headaches trying to install X via the mirror sites).

I've checked the obvious things, permissions on /var/run, on utmp itself,
and the setuid bit on the xterm binary.  Although all of these must be
correct if it's creating the entry in the first place.

Anyone have any thoughts/ideas?

Cheers,

-Paul-

2. printer redirection (to printer on bwlpr)

3. patchadd bug: not cleaning /tmp properly (Sol2.6)

4. PCMCIA Etherlink III howto?

5. UPS Woes (APC Back-UPS Pro)

6. LinuxFromScratch script problem, need help

7. xterms not updating utmp

8. Overmounting /tmp

9. Why does xterm not put in a utmp entry?

10. xterm not updating utmp on exit

11. Xterm not finding /etc/utmp

12. XTerm login not recorded in utmp....