UNIX for Dummies Questions & Answers

I have tried all that. The initial problem was :

even if I kill ALL the admin proccess (everything, inlcuding vuesessions, ttyX, whatever) and do a "w" I still get "admin" logged in.
Doing a "ps -ef |grep admin" after that there is absolutely NOTHING.
I am logged in as root.
I don't want to reboot all the time.
Any ideas please?


Thanks

You wrote:
even if I kill ALL the admin proccess (everything, inlcuding vuesessions, ttyX, whatever) and do a "w" I still get "admin" logged in.
Doing a "ps -ef |grep admin" after that there is absolutely NOTHING.

I can think of couple of things but need more info:

1. If there is no processes associated with the admin userid, then you may just have a bug in the who command and how it is getting information. What OS is this? Have you checked for this as a problem with the OS?

2. What is the LOGIN_AUTO on the admin processes you posted? Is it possible this is a self-spawing processes? How soon after you kill off everything does it come back (the ones with - at the end)

3. Ask yourself - are they doing anything, hurting the system in any way? What is the big deal (especially since you don't see them in a process status (ps -ef). Stop using who - problem solved.

Thehoghunter, admins,

1) The operating system I am using is HP-UX B.10.20 A on 9000/785.

I don't think there is a bug with the "who" command.

2) The LOGIN_AUTO on the admin processes are remote "telnet" sessions to optical multiplexers used for provisioning.

3) The problem is that after a certain number of connected users the system hungs and no more users can log in. This is the reason I am trying to kill any inactive "admin" sessions that migh have left.
People using "exceed" to connect and sometimes (in the case of an error) they just have to close the window down without logging out properly. This leaves inactive sessions at the background without really people being connected. Also these sessions are associated with the appropriate "semaphores" which I have to manually kill all the time. (man ipcs)

Any help will be much appreciated.

I can't find anything on HPUX for this but maybe you have access to HP that I don't (my HPUX Guru is off today).

Check out this article (yes, I know it's not about hpux ). It may give you an idea of what you need to look at especially on the timeouts.

http://www.samag.com/documents/s=1323/sam0110e/

Anymore help concerning the problem above please ??

Cheers !

Here we go again, this is the output of the who -T command.

# who -T|sort
admin + console Feb 26 10:01 old 29754
admin + network Feb 22 16:45 0:01 7409 10.156.81.109:0
admin + network Feb 25 18:31 0:01 25505 10.156.81.68:0
admin + network Feb 26 09:02 0:01 28751 10.156.81.69:0
admin - ttyp5 Feb 26 10:07 0:01 581 preapp01:0.0
admin - ttypd Feb 25 11:57 21:40 20590 192.122.86:0.0
admin - ttyq1 Feb 25 16:54 17:16 24173 192.122.86:5.0
admin - ttyq2 Feb 25 16:54 17:16 24189 192.122.86:5.0
admin - ttyq3 Feb 25 16:54 17:16 24212 192.122.86:5.0
root - pts/0 Feb 26 10:11 . 665 coolp01

How can I kill (kick out) the user admin which resides on the "+ network"

Thank you all for your help. Please see above before you answer.