John, Your comment about a multi-volume data set being OPEN to only one volume at a time is true. However it is fairly simple to have a separate DCB OPEN for each volume so that they can be accessed simultaneously. In addition to allowing customers to be more flexible in specifying space for SORTWKs, it would also allow more concurrent access to your data.
I'm a little surprised that a performance-sensitive product like SYNCSORT does not use a STARTIO driver, in which these issues are mute. I do like your "as needed" design instead of bailing out of the sort; very nice. Tom Harper IMS Utilities Development Team Neon Enterprise Software, Inc. -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Reda, John Sent: Monday, July 02, 2007 7:27 AM To: [email protected] Subject: Re: DFSORT V1R5 - ICE751I and ICE752I Lizette, SyncSort will only use the first volume of a multi-volume SORTWK data set also. Where this becomes less of an issue with SyncSort is that our dynamic allocation of SORTWK data sets is done on an "as needed" basis so if a sort is about to end with an SORT CAPACITY EXCEEDED error it will simply dynamically allocate another SORTWK and processing will continue. The problem with multi-volume SORTWK datasets is that in the output phase the sort needs to have instant access to all of the SORTWK data. A multi-volume data set is OPEN to only one volume at a time. The performance degradation that would be suffered if the sort had to keep switching volumes would be more than anyone would be happy with. John Reda Syncsort, Inc. ---------------------------------------------------------------------- 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
