Nice. I do have a suggestion though, on sessions try "NSCONTROL
KILLSESS=xxxxx" before ABORTJOB. It will fail on serial connections, at
which point you can go ahead and apply ABORTJOB.

The reason for this is even though I'm on 5.5 PP4, I still see
situations where if a user is connected via VT and in a VPLUS screen
and, for example, turns off the PC, applying ABORTJOB instead of
NSCONTROL KILLSESS= can create a ghost session that cannot be removed
anyway other than with a reboot. NSCONTROL KILLSESS=, if tried before
ABORTJOB, seems to work reliably.

John Burke
e-mail: [log in to unmask]


-----Original Message-----
From: Jeff Vance [mailto:[log in to unmask]]
Sent: Friday, July 24, 1998 11:14 PM
To: [log in to unmask]
Subject: ABORTJOB script


Hi all,

A customer was wondering about which job gets aborted when the
"user.acct"
form of the abortjob command is used.  The unfortunate answer is that it
is not easily predictible and, worse, only the first matching job gets
aborted.  Tonight I wrote a simple script that fully supports wildcarded
abortjob's (username, acctname and jobname can be wildcarded) and it
also handles as list of candidate jobs to be aborted.  You need to be on
5.5 express 3 or later to use the script.

I put this script on jazz at:
http://jazz.external.hp.com/src/scripts/index.html

If you discover any problems or have enhancement ideas let me know.

regards,
Jeff Vance, CSY

--