[13126] in Athena Bugs
sun4 7.7K: finger
daemon@ATHENA.MIT.EDU (Christine Ng)
Thu Jan 26 13:46:41 1995
To: bugs@MIT.EDU
Date: Thu, 26 Jan 1995 13:46:31 EST
From: Christine Ng <cstng@MIT.EDU>
System name: george2-sun
Type and version: SPARC/Classic 7.7K (5 update(s) to same version)
Display type: cgthree
What were you trying to do?
[finger cstng]
What's wrong:
[Hi, I get a bug when I finger myself.
I get a segmentation error. OLC has
tried to help me but we couldn't fix
the problem. So they suggested that
I send a bug.
When I finger myself, I get :
Login name: cstng In real life: Christine S Ng
Office: 66-365, 617-253-6521 Home phone: 617-225-9558
Directory: /mit/cstng Segmentation fault (core dumped)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
at the end and a core file is generated.
A copy of that core file has been moved to my
Public Directory. If you don't need it,
please let me know, I'll like to delete it.
I started seeing it yesterday.
Today, when I first logged in, I didn't have
the problem but all of a sudden, it appears
again.
What could be the problem causing this ??
Please look into this.
Christine]
What should have happened:
[No segmentation error.]
Please describe any relevant documentation references:
[ We have modified the following files :
/etc/athena/inetd.conf
and the olc consult is saying that I should be using
a different shell.
I'm using /bin/athena/tcsh but he said I should
be in /bin/csh.
The following dialog with OLC may help.
Topic: unix
Question:
Earlier, I asked a question about having a Segmentation fault
(core dumped) when I finger myself. I logged out and tried again
and the problem appeared to have disappeared so I "done" the
question.
Now, it has come back again.
The last consultant said that the /usr/bin/finger file
may have error. What could be the problem ?
___________________________________________________________
--- Question grabbed by consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 10:59am]
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:10am]
I have a finger program in my directory.
Try the following:
add bdrosen
/mit/bdrosen/sun4bin/finger cstng
Do you have a problem when you do this?
--- User cstng read reply.
[Thu 26-Jan-95 11:10am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:12am]
Yes, same problem as before, I get this at the end :
Login name: cstng In real life: Christine S Ng
Office: 66-365, 617-253-6521 Home phone: 617-225-9558
Directory: /mit/cstng Segmentation fault (core dumped)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:15am]
What shell do you use?
The program dumps core before it gets to you shell entry?
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:15am]
Also, for some reasons, fingering cstng@monk.mit.edu
doesn't show me plan file.
--- User cstng read reply.
[Thu 26-Jan-95 11:15am]
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:17am]
Finger username@monk seems to only give a plan if you are not logged in.
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:18am]
What do you mean what shell do I use ??
Well, we did change the shell line in /etc/athena/inetd.conf from :
shell stream tcp nowait unswitched root /usr/sbin/in.rshd in.rshd
TO :
shell stream tcp nowait unswitched root /etc/rshd rshd
Someone did that so that he could use an x-win emulator
when he telnets from his PC.
After the change, we typed makeserv "something ?? forgot"
and then typed inetd at the promt.
Would that be the problem ?
--- User cstng read reply.
[Thu 26-Jan-95 11:19am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:19am]
No, I've done that million times before. It gives
you a plan anytime.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:20am]
apparently the way monk works has changed.
I am going to see if the changes would cause the problem.
--- User cstng read reply.
[Thu 26-Jan-95 11:20am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:21am]
Well, it's probably problem in my workstation.
I didn't have any core dump when I first logged in
today.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:23am]
What happens if you type:
/usr/ucb/ps aux | grep inetd
--- User cstng read reply.
[Thu 26-Jan-95 11:23am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:24am]
From my directory, I get :
<7>george2-sun:~ _Christine_ >> /usr/ucb/ps aux | grep inetd
root 97253 349 28 ? S Jan 18 0:00 /usr/sbin/inetd -s
cstng 15660 10 150 105 pts/3 S 11:24:18 0:00 grep inetd
root 13922 80 287 60 ? S 14:03:38 0:05 inetd
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:26am]
Ah ah!
you have two inetds running!
--- User cstng read reply.
[Thu 26-Jan-95 11:26am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:28am]
Okay I forgot one thing, we killed if first, I thougth :
athena% kill -HUP `cat /etc/athena/inetd.pid`
athena% inetd
So how do I fix it !
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:28am]
By the say, I'm su-ing root at the moment.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:29am]
as root:
kill the first inetd (inetd -s)
kill 97
i think
then kill -HUP the second inetd
kill -HUP 13922
You don't have to restart it, kill -HUP does it automatically
--- User cstng read reply.
[Thu 26-Jan-95 11:29am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:31am]
Before I do that, one question :
One time I killed -HUP and couldn't telnet to it anymore.
So, we re-ran it again.
Also, can I type the commands from su root or
do I need to login as root ?
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:35am]
So ??? Is it okay to go ahead and kill them ???
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:36am]
You can do it from su
You want to kill the older process, then kill -HUP the newer one.
The older one being the sbin one.
Yep
--- User cstng read reply.
[Thu 26-Jan-95 11:36am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:38am]
I've just done that and now I have only one inetd
<8>george2-sun:~ _Christine_ >> /usr/ucb/ps aux | grep inetd
cstng 15673 12 150 94 pts/3 S 11:36:44 0:00 grep inetd
But I still have the core dump :
Login name: cstng In real life: Christine S Ng
Office: 66-365, 617-253-6521 Home phone: 617-225-9558
Directory: /mit/cstng Segmentation fault (core dumped)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
And I'm unable to telnet to this machine now.
I think I need to rerun by typing inetd again.
Can I do that ???
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:38am]
george2-sun being added to access control list
Trying 18.63.0.124...
telnet: Unable to connect to remote host: Connection refused
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:38am]
type:
/usr/ucb/ps aux | grep inetd
what is returned?
--- User cstng read reply.
[Thu 26-Jan-95 11:39am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:39am]
<11>george2-sun:~ _Christine_ >> /usr/ucb/ps aux | grep inetd
cstng 15684 9 150 97 pts/3 S 11:39:13 0:00 grep inetd
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:39am]
But I have not re-run inetd yet.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:39am]
did you kill both of them or did you kill -HUP one?
--- User cstng read reply.
[Thu 26-Jan-95 11:39am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:40am]
Well, I just typed what you told me :
# kill 97
# kill -HUP 13922
#
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:42am]
It appears that kill -HUP didn't work.
try restarting inetd by.
Make sure you are runnign the one corresponding to your current
inetd.conf.
ie
/etc/athena/inetd
or /etc/inetd
--- User cstng read reply.
[Thu 26-Jan-95 11:42am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:42am]
I'm not sure which one is the current one.
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:43am]
I want the one in etc/athena
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:43am]
probably /etc/athena/inetd
--- User cstng read reply.
[Thu 26-Jan-95 11:43am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:44am]
I'm doing it from the /etc/athena directory
and it says :
# inetd
inetd: couldn't find ISTATE environment variable. Try -s flag.
#
Can I try inetd -s ??
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:44am]
try /etc/athena/inetd
first
--- User cstng read reply.
[Thu 26-Jan-95 11:44am]
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:44am]
If that doesn't work, try the -s flag
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:45am]
Okay, I will, but just want to let you know that I now have
a new process no.
--- User cstng read reply.
[Thu 26-Jan-95 11:45am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:46am]
It worked. But now I have two inetd again :
<16>george2-sun:~ _Christine_ >> /usr/ucb/ps aux | grep inetd
cstng 15701 10 150 93 pts/3 S 11:45:41 0:00 grep inetd
root 15695 79 287 171 ? S 11:45:13 0:00 /etc/athena/inetd
And I still have segmentation fault.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:49am]
do you wnat to try changing back the inetd.conf file?
--- User cstng read reply.
[Thu 26-Jan-95 11:49am]
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:49am]
you only have one inetd now
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:49am]
Yes, I can do that. Since it didn't work anyway.
What do I type after the change ?
--- User cstng read reply.
[Thu 26-Jan-95 11:49am]
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:50am]
after changing it, try:
kill -HUP pid for inetd
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:50am]
Oh, root has one and I have one. Is that what you mean ??
<19>george2-sun:~ _Christine_ >> /usr/ucb/ps aux | grep inetd
cstng 15714 9 150 98 pts/3 S 11:50:00 0:00 grep inetd
root 15695 80 287 155 ? S 11:45:13 0:00 /etc/athena/inetd
So should I still change my inetd.conf file so that
I can fix the segmentation error.
--- User cstng read reply.
[Thu 26-Jan-95 11:50am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:51am]
Do I type kill -HUP pid for inetd , including for inetd ???
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:53am]
No
one is the pid for grep inetd.
(ps aux | grep inetd)
change the /etc/athena/inetd.cong
I am not sure that it will fix anything, but it is worth a try.
No
pid for inetd is a number, specifically the process id number.
so it would be kill -HUP 15695
since that is what ps aux has down for it.
--- User cstng read reply.
[Thu 26-Jan-95 11:53am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:56am]
Okay, I just did that and I still got the segmentation
problem.
Telnet didn't work so I typed /etc/athena/inetd again
But I still have the segmentaiton problem.
Now, there is one inetd for cstng and one for root.
<27>george2-sun:~ _Christine_ >> /usr/ucb/ps aux | grep inetd
cstng 15750 11 150 99 pts/3 S 11:56:18 0:00 grep inetd
root 15738 80 287 171 ? S 11:54:24 0:00 /etc/athena/inetd
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:57am]
Would I have to do a makesev type of command ??
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:57am]
the line:
cstng 15750 11 150 99 pts/3 S 11:56:18 0:00 grep inetd
means that you are running the process grep inetd, NOT inetd.
--- User cstng read reply.
[Thu 26-Jan-95 11:58am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:58am]
I see. There isn't one for root before I rerun inetd
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 11:58am]
I do not know what type of mkserv would affect this.
what shell are you using?
/bin/csh?
--- User cstng read reply.
[Thu 26-Jan-95 11:58am]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 11:59am]
I'm not sure what shell I'm using.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 12:00pm]
try echo $shell
--- User cstng read reply.
[Thu 26-Jan-95 12:00pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 12:00pm]
george2-sun# echo $shell
/bin/athena/tcsh
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 12:01pm]
What does more /etc/passwd.local | grep cstng retrun?
--- User cstng read reply.
[Thu 26-Jan-95 12:01pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 12:02pm]
george2-sun# more /etc/passwd.local | grep cstng
cstng:*:823:101:Christine S Ng,,66-365,6172536521,6172259558:/mit/cstng:/bin/csh
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 12:03pm]
other people are also using the same shell but
they don't have problems. But they don't use it
as much as I do.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 12:03pm]
that might be your problem ....
you should be running /bin/csh but are running /bin/athena/tcsh ...
type ls -l /bin/csh
--- User cstng read reply.
[Thu 26-Jan-95 12:03pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 12:03pm]
george2-sun# ls -l /bin/csh
lrwxrwxrwx 1 root root 16 Oct 3 12:46 /bin/csh -> /bin/athena/tcsh
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 12:05pm]
Ok that explains why you are using tcsh instead of csh
--- User cstng read reply.
[Thu 26-Jan-95 12:05pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 12:05pm]
So how can I fix it ???
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 12:06pm]
I am still not sure as to what is causing the problem here.
I am trying to figure out what would cause this.
Do you have this problem on other machines?
--- User cstng read reply.
[Thu 26-Jan-95 12:06pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 12:07pm]
Well, as I said, when I first logged in, I didn't
have the problem. But after I while, it appears
again. I did use kinit -l eariler to get a longer
ticket.
As for other workstation, I have not seen the same
probelm before.
In fact, I just started seeing this yesterday,
I think. I never have any problem before, I think.
*** Reply from consultant bdrosen@AFTERBURNER.MIT.EDU [0].
[Thu 26-Jan-95 12:14pm]
I am really not sure what could be causing this at all.
Maybe another consultant might have run across this before.
kinit should not effect it.
I have to go now but I will forward you to the next available consultant.
If you are done with your question, please type done at the OLC prompt.
--- User cstng read reply.
[Thu 26-Jan-95 12:14pm]
--- Question forwarded by bdrosen to pending.
[Thu 26-Jan-95 12:24pm]
--- Question grabbed by consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:17pm]
*** Reply from consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:20pm]
If you can reproduce this, send a bug report (with "sendbug").
The quickest way to see why it is happening is to take the core
dunp and get a backtrace. The person who does this will want a
copy of the source code for finger.
It is most likely a problem with some field in your posswd entry
being wrong, not having a delimiter, etc. You could have a finger
that is compiled with the wrong system library and has the wrong
passwd structure, for example.
No UNIX program should ever core dump on you, it should just tell
you whqt is wrong and clean up after the mess. So it is a bug,
regardless.
--- User cstng read reply.
[Thu 26-Jan-95 1:21pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 1:22pm]
Okay, how do I send my core to that person ?
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 1:26pm]
Never mind. I've moved it to my Public directroy.
Would reentering my finger info help ???
I've already changed my password.
*** Reply from consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:26pm]
Wait until they ask. They may not need you to produce one, and will
probably want to make sure it is produced by a binary with a complete
symbol table.
--- User cstng read reply.
[Thu 26-Jan-95 1:26pm]
*** Reply from consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:26pm]
Heck, if you start changing things, we may lose the ability to
reproduce the bug; why would you do that?
--- User cstng read reply.
[Thu 26-Jan-95 1:27pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 1:27pm]
You said maybe I have an error in my file.
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 1:30pm]
So should I "done" the question or should I leave it here.
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 1:32pm]
Did you say :
sendbug ??
The mail was returned.
----- Transcript of session follows -----
Connected to SOUTH-STATION-ANNEX.MIT.EDU:
>>> RCPT To:<sendbug>
<<< 550 <sendbug>... User unknown: Bad file number
550 sendbug... User unknown
*** Reply from consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:35pm]
Did you send the bug report?
I do not see where I said any of your data was in error, when I say
"your passwd entry" I mean the system file on the machine you are
currently using, not any file owned by you. If you have been editing
things as root, be sure that you have not changed the format of the
entry (added extra colons, left out fields, etc.).
I don't think you'll get more help from olc, so you can mark it done
if you like.
If you continue to try and solve it yourself, go ahead and run finger
under a debugger, and see what it is doing when it take the seg fault.
Most likely it will be doing something with a string, under some assumption
as to length or size that does not hold in your case -- but it could be
almost anything.
--- User cstng read reply.
[Thu 26-Jan-95 1:36pm]
*** Reply from user cstng@GEORGE2-SUN.MIT.EDU [0].
[Thu 26-Jan-95 1:37pm]
But sendbug@mit.edu does not exist .
Where should I send the report to.
*** Reply from consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:38pm]
sendbug is a program to execute, not a mail address.
--- User cstng read reply.
[Thu 26-Jan-95 1:38pm]
*** Reply from consultant yba@THE-GOOD.MIT.EDU [0].
[Thu 26-Jan-95 1:42pm]
I need to log out. Will forward you for now.
--- Question forwarded by yba to pending.
[Thu 26-Jan-95 1:42pm]
--- User cstng read reply.
[Thu 26-Jan-95 1:43pm]
]
