Re: Mix 3390-3 and 3390-9

New Message Reply Date view Thread view Subject view Author view Other groups

Subject: Re: Mix 3390-3 and 3390-9
From: Gustavo Torres (clgtorres@gmail.com)
Date: Mon Aug 11 2008 - 10:42:38 EDT 

You could apply all PTF's of the list "VSE/VSAM 6.6 APARs for VSE/ESA V2.6"

http://www-03.ibm.com/servers/eserver/zseries/zvse/support/vsamolder.html#vseesa26


On 8/11/08, Edward M. Martin <EMartin@aultman.com> wrote:
> Hello Ronald,
>
>        We do have one problem with 3390-9 VSAM.
>
>        We are on VSE 2.6.1 with every fix I can find for VSAM under
> z/VM on DS6800 split into 3390-3 and 3390-9.  And this problem happened
> on the EMC Symmetric too.
>
>        Problem.   Implicit defined files going to a second volume on
> 3390 mod 9 (only) can not be read.  The Last Block of the first extent
> on the second volume is junk.  The file is written out with our error.
> It is upon the read of the file that cause the permanent I/O error.
>
>        1) Has to 3390 mod 9.   3390 mod 3 does not have the problem.
>        2) Has to be implicit define.  Explicit define works.
>        3) and PHYREC-SIZE--------10240.
>
>        Because there is nothing on IBM and we are out of support I have
> not been able to correct the implicit problem.
>
>        Corrections.
>
>        1) Change the RECORDS or RECSIZE to ensure the file stays on one
> volume.                 Easiest but may not keep to one volume
>        2) Explicit define.             Always works.
>        3) Change the Block size to get a different PHYREC-SIZE.
>                                                Programmers do not like.
>        4) Change to ESDS.              Programmers do not like.
>
>        5) Install z/VSE 4.1.2.  Programmers like and I am doing.
>
>        Other than this one problem,
> I like the 3390-9 devices.
> We have 8-9 volumes per catalog instead of 24-27 volumes.
> z/VM VSE/VSAM is a worry any more (max of 26 volumes).
>
> Easier to manage.
>
> Performance is better than what I expected but that is not related to
> the device type but to the 4 FICON channels versus the 6 ESCON we had
> before.
>
> Our batch cycle time decreased from 6 hours to 4.5 hours with out any
> jcl changes.  I have dropped another .5 so far and am working on other
> ways to eliminate the bottleneck of tapes.
>
> Ed Martin

New Message Reply Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b25 : Sat Aug 30 2008 - 00:35:13 EDT