home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!umn.edu!msus1.msus.edu!stafford.winona.msus.edu!user
- Newsgroups: vmsnet.internals
- Subject: Re: re:... idle terminals
- Message-ID: <Stafford-101192114012@stafford.winona.msus.edu>
- From: Stafford@Vax2.Winona.MSUS.EDU (John Stafford)
- Date: 10 Nov 92 11:50:27 -0600
- Followup-To: vmsnet.internals
- References: <9055342@MVB.SAIC.COM>
- Distribution: world
- Organization: MSUS
- Nntp-Posting-Host: stafford.winona.msus.edu
- Lines: 26
-
- In article <9055342@MVB.SAIC.COM>, "Bob George" <BOB@aeneas.ims.disa.mil>
- wrote:
- > Here's a stupid comment about just being killed by an idle terminal
- > monitor. Aren't there a few flavors of them that are on the internet
- > along with a commercial one or two that allows you to configure them
- > to ignore certain usernames/terminals/... Get one of them???
-
- If FORCEX is not an issue, I cannot see why this is such a complex
- problem. We have an elementary DCL procedure servering as a good
- idle process killer. It can make execeptions based upon username,
- group, terminal, job and kill from the top down, protect system jobs, have
- on-off hour times for certain kills, most of the good stuff.
-
- Elementary
-
- create Indexed file
- Loop:
- wait n unless first iteration (the max idle time)
- Get PID and process info
- Process exemptions
- Append padded USERNAME to F$TIME()
- Check Indexed file - does this key exist?
- if it does then the job has been idle for n minutes.
- KILL it.
- if it does not exist, then Put this as a KEY to Indexed file
- repeat Loop
-