> -----Ursprungligt meddelande-----
> Från: IBM Mainframe Discussion List [mailto:ibm-m...@bama.ua.edu] För Paul
> Gilmartin
> Skickat: den 2 juli 2010 15:40
> Till: IBM-MAIN@bama.ua.edu
> Ämne: Re: SV: Delete all members of a PDS that is allocated
> 
> On Fri, 2 Jul 2010 13:26:41 +0200, Thomas Berg wrote:
> >> >
> >> >How ?  The "shut the door" job is *waiting* for EXCL ENQ (OLD).
> >> >
> >> As I read it, "shut the door" never completes its wait.  So
> >> you don't know how many other jobs may have previously
> >> allocated the data set SHR, opened it, done BLDLs, and not
> >> yet closed it.  They are holding TTRs which will point to
> >> unpredictable content once you compress (w DISP=SHR).
> >
> >(The "STD" job is cancelled by the compress job when it's finished.)
> >Yes, a couple of other jobs has opened the pds and if they haven't
> >finished their update it may theoretically cause local corruption
> >in the pds.  But I have never experienced that, maybe because of the
> >30 seconds wait I do before beginning the compress.
> >
> I was concerned less with directory corruption than with the harmful
> effect on other DISP=SHR jobs already running.
> 
> >And of course, this is in a test environment.
> >
> Ah!  So you don't care if a few jobs fail.  I, also, often operate
> in that mode.

Well, I do care, but it's not a very big problem if it happens and, 
as mentioned, in practice it very seldom do.
 
> >BTW, the "STD" is not a problem in this case, the only potential
> >problem is the compress (job).
> >And if You are worried You can always wait until all jobs that may
> >have opened the pds are finished before You start the compress. Then
> >there is no problem anymore regarding the risk ds corruption.
> >
> IOW, simply submit a compress job with DISP=OLD.  I sometimes do
> that, too.  Then the operators telephone me to complain about the
> flood of MIM ENQ messages on the console, and to relay complaints
> from other testers attempting to allocate the data set.

The problem with the DISP=OLD compress job is that some jobs that 
enques the pds could be running VERY long time after the access to 
the pds due to unrelated (but valid/correct) causes. 
(And with DISP=OLD You are placed behind all other jobs that are 
*waiting* for the enq.) 


 
Regards, 
Thomas Berg 
_________________________________________ 
Thomas Berg   Specialist   A M   SWEDBANK 

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to