[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

immortal login



I've a problem with the PAM-ized version of (util-linux's) login.

I've a server that accepts dial-in connections and often
I find lots of login processes still alive with no children
(after the connections have been closed)

I think that either the login processes are still asking for username/
password (but there shouldn't be a timeout???) or they don't die 
when their children do.

here his some output of "ps -axfu"
USER       PID %CPU %MEM  SIZE   RSS TTY STAT START   TIME COMMAND
....
root       178  0.0  0.5   884   356  ?  S   19:31   0:00 inetd 
root      4215  0.0  0.7  1144   488  ?  S   22:20   0:00  \_ in.telnetd 
root      4216  0.0  0.8  1036   568  p4 S   22:20   0:00  |   \_ /bin/login -h
root      4294  0.0  0.7  1144   488  ?  S   22:22   0:00  \_ in.telnetd 
root      4295  0.0  0.8  1036   568  p9 S   22:22   0:00  |   \_ /bin/login -h
root      4313  0.0  0.7  1144   488  ?  S   22:25   0:00  \_ in.telnetd 
root      4314  0.0  0.7  1028   500  p3 S   22:25   0:00  |   \_ /bin/login -h
root      4589  0.0  0.7  1144   488  ?  S   22:35   0:00  \_ in.telnetd 
root      4590  0.0  0.8  1032   536  pa S   22:35   0:00  |   \_ /bin/login -h
root      7184  0.0  0.7  1144   488  ?  S   01:09   0:00  \_ in.telnetd 
root      7186  0.0  0.8  1036   568  p6 S   01:09   0:00  |   \_ /bin/login -h
root      7201  0.0  0.7  1144   488  ?  S   01:12   0:00  \_ in.telnetd 
root      7202  0.0  0.9  1036   576  p7 S   01:12   0:00  |   \_ /bin/login -h
root      7395  0.0  0.7  1144   488  ?  S   01:18   0:00  \_ in.telnetd 
root      7396  0.0  0.8  1032   544  p8 S   01:18   0:00  |   \_ /bin/login -h
root     10757  0.0  0.8  1144   544  ?  S   09:29   0:00  \_ in.telnetd 
bertacco 10758  0.0  1.1  1036   752  p2 S   09:29   0:00  |   \_ /bin/login -h
bertacco 10759  0.0  1.1  1260   712  p2 S   09:29   0:00  |       \_ -bash 
....

there are lots of logins with root uid so probably they don't have completed
authentication (not sure about this) and are still prompting the user. 
Anyway some of them have been running for more then 10 hours. Usually they
should timeout long before...

There is also a user "bertacco" with a bash. Actually I can see that he is
not really connected (the modem is on-hook); so probably my 
portserver doesn't close the telnet session correctly when the
modem disconnects...(but I think logins with no children should die anyway).

I'm using RedHat 4.1 with all official patches applied. I didn't have this problem
when I was using slackware 3.0 (same portserver).
Before trying to recompile login.c with no PAM and see what happens,
I'd like to know if you have some suggestions. (maybe PAM doesn't
matter at all)

Thanks,
  Livio Bertacco

P.S. Is it really necessary to have login fork-ing before exec-uting the shell?



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index] []