[11508] in Athena Bugs
sun4 7.6J: codecenter
daemon@ATHENA.MIT.EDU (alexp@MIT.EDU)
Wed Dec 8 16:15:23 1993
From: alexp@MIT.EDU
Date: Wed, 8 Dec 93 16:14:43 EST
To: Calvin Clark <ckclark@MIT.EDU>, bugs@MIT.EDU
Cc: alexp@MIT.EDU
Calvin,
We have received the following response from Centerline to your bug report:
******************************************************************************
You can use the command: killms
That should (hopefully) kill off all the appropriate processes. I've
attached yet another document in case killms doesn't work. Ultimately
(worse case) you have to do kill them by hand. But let's see if the
killms command works for you. I'll leave the incident open till you can
let me know if this does/does not help.
====================Document 191.32====================
Short Description:
Sometimes CenterLine/<arch-os>/bin/killms doesn't kill everything
Try
FOR SUN % ps -auxww | grep clpm
FOR HP % ps -edaf | grep clpm
then % kill -USR2 <clpm-pid>
Long Description:
>I had a session on an X display which was hung so I did a killms; then
>clms_query showed up no active sessions. However, the processes
>"cx.strip" and "clpm" survived and suddenly became very expensive. Should
>they have been removed by the killms action?
Things that might cause killms to fail to kill all CenterLine process:
o Hanging clpm
o Hung or crashed or killed clms
WORKAROUND
----------
% kill -USR2 <clpm-pid>
This tells clpm to shut everything down.
If that doesn't work, kill remaining processes by hand.
******************************************************************************
This is probably as much of a solution as we are going to get. We don't have
any other documentation on killms, but it is in
/mit/codecenter4/CenterLine/sparc-solaris2/bin
Tyoing killms -h will give you a small additional amount of information. I
hope this is of some use in case you run into this again.
Alex Prengel, Jr.
ACS Software Acquisition Coordinator
>From ckclark@Athena.MIT.EDU Fri Sep 3 20:52:39 1993
>Received: by cascade.MIT.EDU (5.57/4.7) id AA13715; Fri, 3 Sep 93 20:52:38 -04>00
>Received: from W20-575-120.MIT.EDU by Athena.MIT.EDU with SMTP
> id AA22485; Fri, 3 Sep 93 20:52:32 EDT
>Received: by w20-575-120.MIT.EDU (5.0/4.7) id AA19210; Fri, 3 Sep 93 20:52:23 >EDT
>Date: Fri, 3 Sep 93 20:52:23 EDT
>Message-Id: <9309040052.AA19210@w20-575-120.MIT.EDU>
>From: Calvin Clark <ckclark@MIT.EDU>
>To: bugs@Athena.MIT.EDU
>Subject: sun4 7.6J: codecenter
>Content-Length: 1280
>
>-------
>System name: w20-575-120
>Type and version: SPARC/Classic 7.6J
>Display type: cgthree
>
>What were you trying to do?
>
> Run codecenter out of the codcenter4 locker. I did:
>
> add codecenter4
> codecenter -motif &
>
>What's wrong:
>
> It brought up a window with the following messages:
>
> cllmgr: cannot initialize license system, status = 1
> >>> license_init: unable to locate license file
>
> cllmgr: cannot initialize license system, status = 1
> >>> license_init: unable to locate license file
>
> Then, about 3 minutes later, I got a light blue window
> telling me: ``Initializing windows. Please wait.''
>
> I went to the bathroom. When I came back a few minutes
> later, nothing had happened. I observed that there
> were about a dozen or so processes running out of the
> codecenter4 locker. The blue window was hung (it wasn't
> refreshing on expose events, and the hourglass cursor
> was still active.)
>
> This is repeatable.
>
>What should have happened:
>
> Even if the license server is down, it shouldn't hose
> my machine. Haven't they heard of timeouts? It should
> display some informative error diagnostic and either
> exit or allow me to to quit. I shouldn't have to kill
> off n random processes manually.
>
>Please describe any relevant documentation references:
>
> N/A
>
>NETPROB-NOTE:
>Note: carla Fri Sep 17 12:39:33 1993
>Forwarding ti f_l, since 3rd party sw.
>------
