David, I checked the SYS1.PARMLIB. In the example they are using (see below) 2048 DEFINE CLUSTER (NAME(?UID.MCDS) VOLUMES(?MCDSVOL) - CYLINDERS(?CDSSIZE) FILE(HSMMCDS) - STORCLAS(?SCCDSNM) - MGMTCLAS(?MCDFHSM) - RECORDSIZE(435 2040) FREESPACE(0 0) - INDEXED KEYS(44 0) SHAREOPTIONS(3 3) - SPEED BUFFERSPACE(530432) - UNIQUE NOWRITECHECK) - DATA(NAME(?UID.MCDS.DATA) - CONTROLINTERVALSIZE(12288)) - INDEX(NAME(?UID.MCDS.INDEX) - CONTROLINTERVALSIZE(2048))
--- On Wed, 5/6/09, O'Brien, David W. (NIH/CIT) [C] <[email protected]> wrote: From: O'Brien, David W. (NIH/CIT) [C] <[email protected]> Subject: Re: DEFINE CLUSTER MYSTERY To: [email protected] Received: Wednesday, May 6, 2009, 1:15 AM How about IBM's opinion? The folks who wrote the code should know the best CI sizes. Refer to SYS1.SAMPLIB(ARCSTRST) Then do a find for NAME(?UID.MCDS) You'll see that IBM uses a Data CI of 12K If your management is allowing auditors to dictate VSAM parameters then you have my sympathy. Dave O'Brien NIH Contractor ________________________________________ From: IBM Mainframe Discussion List [[email protected]] On Behalf Of esmie moo [[email protected]] Sent: Tuesday, May 05, 2009 9:13 AM To: [email protected] Subject: Re: DEFINE CLUSTER MYSTERY Our auditors don't want to make changes if it "ain't broke". I tried explaining the situation but I am receiving static. Thanks for the advice. Maybe, they will accept another person's opinion. Many thanks. --- On Wed, 5/6/09, O'Brien, David W. (NIH/CIT) [C] <[email protected]> wrote: From: O'Brien, David W. (NIH/CIT) [C] <[email protected]> Subject: Re: DEFINE CLUSTER MYSTERY To: [email protected] Received: Wednesday, May 6, 2009, 12:56 AM NO, but why would you want to use a 4K Data CI? Dave O'Brien NIH Contractor ________________________________________ From: IBM Mainframe Discussion List [[email protected]] On Behalf Of esmie moo [[email protected]] Sent: Tuesday, May 05, 2009 8:54 AM To: [email protected] Subject: Re: DEFINE CLUSTER MYSTERY David, Thanks for the advice. If I use the value which VSAM substitutes would there be a problem with DFHSM? --- On Wed, 5/6/09, O'Brien, David W. (NIH/CIT) [C] <[email protected]> wrote: From: O'Brien, David W. (NIH/CIT) [C] <[email protected]> Subject: Re: DEFINE CLUSTER MYSTERY To: [email protected] Received: Wednesday, May 6, 2009, 12:29 AM Try 18K for your Data CI. The Index CI contains pointers to each Data CI. Using a 4K Data CI will produce 180 Data CI per CA. A 2k Index CI is not large enough therefore VSAM corrects your error. Dave O'Brien NIH Contractor ________________________________________ __________________________________________________________________ Looking for the perfect gift? Give the gift of Flickr! http://www.flickr.com/gift/ ---------------------------------------------------------------------- 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
