On Fri, 7 Oct 2005 11:24:38 -0600, Lance Kopplin wrote: >----- Original Message ----- >From: "Mark Steely" <[EMAIL PROTECTED]> >> We are z/OS v1r4. We have a started task which creates a GDG file. This >> started task stays ups unless it abends or an IPL happens. This GDG is >> SMS controlled. There are several files which will not migrate because >> the GDG index is enqueued. Is there a way to get around this so the >> files will be migrated. > >We do something very similar to this. We have a started task that reads >syslog off spool and writes it into a GDG, beginning a new GDG at midnight >each night. The way this is coded is to do a dynamic allocation of the +1 >generation, which would be exclusive and should ENQ on the GDG base. After >the dataset is actually created, it is then deallocated (via dynamic >allocation) and then allocated again with DISP=SHR. The DISP=SHR does not >ENQ on the GDG base.
So you use the first dynamic allocation to create the dataset, then deallocate it (freeing the exclusive ENQ) and then reallocate the same generation as SHR in order to avoid the base ENQ? Shrewd. >Then the output dataset, allocated DISP=SHR as it is, is opened for output >with EXTEND and data is written to it every minute or so. So it ends up >being one writer and many readers. I've often wondered why nobody ever >sees garbage at the end of the dataset, but it doesn't happen. Probably >using up all my luck that doesn't show up in the lottery. Lance, correct me if I don't understand, but are you then running this started task all day (and night) so that your syslog GDG is more-or- less "up to the minute" with the JES syslog buffers? And you have multiple GDG syslog readers throughout the day while you are writing to it with that started task? -- Tom Schmidt Madison, WI ---------------------------------------------------------------------- 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
