John Thanks for those comments. We mostly use DFSORT on other accounts and my immediate take on the problem (while waiting fro proper analysis) was that it was like the Memory Object sort problems that occurred for a while with DFSORT but I had forgotten the VLSHRT fix. Will look at whether that could be happening here.
regards, Anthony Fletcher - NZ MIITP Team Lead NZ SMM (AirNZ, Westpac NZ , TelstraClear NZ and NWM AU) IBM Strategic Outsourcing Delivery Server Systems Operations Server Management Mainframe Mainframe Software Program Manager NZ z/OS Technical Lead A/NZ Ph: Direct +64 4 576 8142, tieline 61 929 8142, ITN *869298142, mobile +64 21 464 864, Fax +64 4 576 5808. Internet: [email protected], Sametime: [email protected] "The biggest threat to effective communication is the belief that it has occurred" "Winners make commitments, Losers make promises" From: "McKown, John" <[email protected]> To: [email protected], Date: 04/12/2012 10:04 Subject: Re: SYNCSORT Sent by: IBM Mainframe Discussion List <[email protected]> Sheer guess on my part. Take it for what it's worth. We don't use SYNCSORT. But I've seen some weird SOC4-10 abends in the past. Is the input file variable length records? Does SYNCSORT have something equivalent to the VLSHRT parameter? Anyway, what could be happening is that one or more of the records is "too short" for your control fields. Sometime during the sort, one of these "too short" records ends up being placed so that the last byte of the record is on the last byte of a 4K page, and the subsequent virtual page is invalid. Since the SORT control field is too long, the SORT program does a compare which causes the comparand field in the record to extend into the invalid address range. Which causes an S0C4-10 abend. If SYNCSORT is like DFSORT, it "compiles" the SORT control fields into instructions in an STORAGE OBTAIN'd area, and so they are not part of any executable module. So the abend formatter ends up saying the abend occurred in an unknown module. DFSORT avoids this problem with VLSHRT by padding any records which are "too short" with 0x00 bytes at the end so that they will not compare high (might be low or equal). -- John McKown Systems Engineer IV IT Administrative Services Group HealthMarkets® 9151 Boulevard 26 • N. Richland Hills • TX 76010 (817) 255-3225 phone • [email protected] • www.HealthMarkets.com Confidentiality Notice: This e-mail message may contain confidential or proprietary information. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message. HealthMarkets® is the brand name for products underwritten and issued by the insurance subsidiaries of HealthMarkets, Inc. –The Chesapeake Life Insurance Company®, Mid-West National Life Insurance Company of TennesseeSM and The MEGA Life and Health Insurance Company.SM > -----Original Message----- > From: IBM Mainframe Discussion List [mailto:[email protected]] > On Behalf Of Anthony Fletcher > Sent: Monday, December 03, 2012 2:47 PM > To: [email protected] > Subject: SYNCSORT > > Any SYNCSORT users out there? We have one job that gets an S0C4-10. The > problem can be reproduced with a three line set of data, so its not a > capacity problem. There are two sort fields. If the second one's > displacement is reduced the problem goes away but it otherwise looks > like an addressing problem. The sort length plus displacement is > nowhere near as long as the LRECL of the data set. > The abend does not appear to be in any SYNCSORT module. > It is version 1.3.0 running on z/OS 1.13. > Waiting for vendor agent to get back. > Anyone experienced anything like this? > > ---------------------------------------------------------------------- > 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 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
