Well, the original LRECL is 304. Let's say every record is 300 bytes. Fills up 
the block, writes it. No room on the track for the next block. Next block will 
be on a different track. Wasteful.

I mention it because these days it seems that many people (short on experience) 
think "I've got a lot of data, what's the maximum record-size I can use", so 
they reach for a manual, discover, and happily "design" their file to have 
32000+byte records.

Yes, I have no clue if "waste" occurs on the underlying physical media, but 
that's not the point. Or is it? I don't know. 

I'd like to know if it doesn't matter, as then I can stop raising it every time 
I see it. Or can I? When I get to thinking of VSAM, things go Klunk! again. Or 
do they...?


On Friday, 4 March 2016 14:56:31 UTC, Paul Gilmartin  wrote:
> On Fri, 4 Mar 2016 04:58:45 -0600, Bill Woodger wrote:
> 
> >Wouldn't that be a little wasteful? One block per track.
> >
> >On Friday, 4 March 2016 10:48:40 UTC, Steve Coalbran  wrote:
> >> ALLOC with VB LRECL=32756 ?
> >>  ;-)
> >>
> It depends on how clever QSAM is at using the balance of a block which
> can contain several short records.  The worst case is one record per
> track.
> 
> It has been discussed here that Binder will write short blocks in order
> to fill a track; QSAM generally will not.
> 
> -- gil
> 
> ----------------------------------------------------------------------
> For IBM-MAIN subscribe / signoff / archive access instructions,
> send email to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to