On Wed, 8 Apr 2009 07:56:01 -0500, Staller, Allan wrote: >The use of IEFUSI is intended to be a trigger. This is to force invoke >IEFUTL at an interval so that elapsed time can be calculated. If the >elapsed time calc in IEFUTL indicates, for example, 10 miniutes elapsed, >set the IEFUTL retcode to extend the job. If it indicates 2 hrs & 1 >minute elapsed, allow the job to abend. > >This will not be precise, but it will be close. Feel free to contact me >off list if you like.
It will not be close unless you make the CPU time limit very small. Certainly if you allow 5 minutes CPU time it will not be within one minute of wall clock time. It *especially* will not be close if the system is heavily loaded and the test job in question is not getting much access to the CPU. Perhaps if you make the CPU time limit 1 second it will be close enough, but then you will have a lot of IEFUTL exit calls. Sounds like an expensive solution to me. -- Tom Marchant > >HTH, > ><snip> >> Code/install IEFUSI to limit *ALL* test jobs to (e.g.) 5 min of cpu >> time. Code/install IEFUTL to calc elapsed time and extend (or not) as >> required. IIRC, All of the required info needed should be available to >> the exit >> >> Adjust CPU time limit to desired granularity based on your shop. Not >too >> precise but it should come close to your needs. >> >> HTH, >> > >Problem: I must restict WALL CLOCK or ELAPSED time, not CPU used time. >It's a weird political thing. ></snip> > >---------------------------------------------------------------------- >For IBM-MAIN subscribe / signoff / archive access instructions, >send email to [email protected] with the message: GET IBM-MAIN INFO >Search the archives at http://bama.ua.edu/archives/ibm-main.html ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
