In article <[log in to unmask]>, [log in to unmask]
says...
>
>In article <[log in to unmask]> "Michael T. Kapsak"
<[log in to unmask]> writes:
>>From: "Michael T. Kapsak" <[log in to unmask]>
>>Subject: Aborting sessions
>>Lines: 39
>
>
>>We are running a Series 967 with two a manufacturing application
>>written in Fortran and a Financial app in RPG. Bot hare
>>maintained by a company called Computer Financial Services in
>>Washington state.
>
>>My problem, for at least the past 7 years, has been to come up
>>with a guaranteed way to make sure sessions are aborted prior to
>>our nightly production backups. In the old days there were
>>operators to take care of sessions prior to backing up, but since
>>we now have only the 5:00am - 4:00pm coverage, I have to rely on
>>a logoff job using MPEX and attempting ABORTJOB's and ABORIO's.
>>It does not always get rid of the sessions I want. They either
>>end up as ghost sessions or the HP has determined that the
>>process state is critical enough not to abort it.
>
>We have a 957RX with terminals/PC's connected to DTC's via direct connect,
>mux's, modems and router/ISDN/bridge networks.
>
>We are running MPE/iX 4 and have been sucessfull with the following
combination
>of MPE commands in a sheduled job (we dont have MPEX) and an in-house
program
>written in BASIC/V.
>
>from job .......
>:JOBSECURITY LOW
>from program ...
>:ABORTIO  ldevno
>:ABORTJOB jobnum
>:ABORTIO  ldevno
>
>The last ABORTIO seems to pick up even the most stubborn jobs.
>
>Hope this helps in your situation.
>
 
Hello,
 
this does NOT allways work, exspecially with virtual terminals
over a pc network, or via x25.
 
Try this
 
NSCONTROL KILLSESS=#s1234
 
I found this allways works, also in situations where abortio and
abortjob don't work.
(i saw a CE do this once, and typed listredo afterwards, I'm not shure if
this is really supported)