Minutes of the GSE UK IMS Working Group Meeting - IBM South Bank, 4^th December 1997

Chairman: Ian Tyson sitting in for Frank Fleming (GBBCO4CB @ IBMMAIL , frank.fleming@barclays.co.uk) - 01565-613581

Secretary: Peter Armstrong (p_armstrong@bmc.com) - 01276-419835

Administration

• Thanks
  • Welcome to South Bank – nice to see the room full (unlike the DBCTL meeting 2 weeks ago, which was very poorly attended). Apologies from Frank, who is over in America with IBM playing with IMS. Thanks to Ian for organising. We need volunteers for presentations next year please. Major item on the agenda today is IMS Version 6, including a live demo!

• Minutes
  • Not everybody received minutes – apologies for this, which was caused by an unreadable floppy (Don't you just love PCs????!!!!). Apologies to anyone who didn't get the bits they should have. Please check the signing list as it goes round and update as necessary. Thanks to Barclays for all the photocopying this time. Andrew Jack has now moved to Standard Life, who have kindly agreed to let Andrew carry on sending out the minutes. Please try and print your foils many to a page to save the amount of copying we have to do.

  • Should we be sending out the minutes electronically?

• Matters arising
  • None.

• New Members / attendees
  • Welcome to Chris McCanna representing J&H Marsh McLennan Services Ltd, and Hadrian Thackray from ITSA.

• Correspondence
  • None.

• Apologies
  • List attached.

• Other matters
  • UK Region Board Meeting – chairman is now Phil Howard-Knight, to whom we hoped to present a little trinket today, but he couldn't make it. We still have some attendees who are not members of GSE. Please check that you are members and have paid. Web site is www.gse.org. GSE UK will be setting up a proper site soon.

  • Next European meeting is in May in Gothenburg – please submit papers if you are working on something interesting.

  • GSE are looking at closer co-operation with CUA in future. In fact the May meeting will be a joint meeting.

  • Book from Circle group in Brighton with all their education courses.

IBM What's New - Alan Cooper (IBM ISC)

1. Glen Galler has returned to Santa Teresa, and is now replaced by Tom Ramey, who is presenting this afternoon.
2. Details of education classes. Version 6 class dates will be announced soon. If you want any of the "scheduled on demand" courses contact IBM education or Alan/Pete.
3. IBM Technical Conference will be in Munich in June, joint with DB2.
4. IMS Version 6 GA will be 26/12/97, including PRPQ to provide support for CICS local DL/1 (i.e. pseudo Version 4 of IMS). End of support for Version 4 otherwise is 31/12/98. New flyer on Version 6 made available.
5. Replacement product for IMSPARS/IMSASAP – the IMS Performance Reporter (5697-B89). Flyer available.
6. DBICF V6 (5697-B92) will be available at V6 GA, with support for V6 (and only V6). Flyer available.
7. New release of Message Requeuer available with support for shared message queues. Flyers for this and other things like Index Builder, Partitioned DB, IMS C/S Builder, Internet connection, Objects, customer experiences etc. available.
8. Java for OS/390 available on the Internet. Provides access to IMS, CICS, DB2, MQ.
9. Details of all the latest and greatest red books.

• A copy of the foils is enclosed.

Where can I recover to nowadays in IMS? (Peter Armstrong - BMC)

• Peter did some testing of IMS in an n-way sharing sysplex earlier this year and discovered that there are several scenarios where you cannot recover data bases with the standard IMS utilities!! He described the problem, showed the IBM response, and of course gave the good news that the BMC products don't have the same problem.

• He then went on to show that there are some serious problems to be overcome if you want to design a disaster recovery plan in an n-way sharing environment. This is due to the fact there is no point you can recover to unless you close the system down, /DBR databases regularly. Peter explained the problem and described how BMC are attacking this problem by providing the ability to recover to any point in time (even in the middle of a log). This problem also applies if you are running IMS and DB2 together.

• He is about to run another set of tests in a non-sharing environment, where there are some similar strange problems. See also under AOB for the APAR/PTF numbers for various problems raised here and at the last meeting.

• Peter recommended keeping an eye on recent DBRC/DSSN maintenance (see numbers below), and always sorting logs prior to recovery (preferably using a product with a plus on the end!).

• A copy of the foils is enclosed.

Indexices – Discussion topic

• Discussion on calculating VSAM parameters for index databases. Have to carefully calculate CI and CA sizes in order to use space efficiently. There is a spreadsheet on Dougie's website, which will give you all the bits you need for calculating these. Look at the VSAM books as well – this is a VSAM KSDS issue.

Problems forum, hints and tips, contacts for lunch

• Requirements – none.
• Web address for hints and tips is web.online.co.uk/dougie.lawson/gse_files
• Hints and tips
  • Name: Dougie Lawson - IBM

    Problem: How are we going to handle the EURO? What symbol X'9F', #, £ with another slash, ¤ ? Where is on the keyboard? (Question prompted by an Irish customer).

    Contacts/Comments: Put question on Forum – try the printer people as they deal with code pages.

  • Name: Alan Lovett – ITSA – 01253-688311

    Problem: 2 problems with Data Propagator / Non-relational.

    PN90143/UN97389 - GENJCL.USER %SELECT SLDS skips valid logs, resulting in DProp selector failure. This is a DBRC APAR. The PTF was not marked HIPER, but now is.

    New PMR (no APAR number at time of writing). Similar to 1), but for compressed log entries. Selector cannot account for gaps created by compressed logs (GENJCL.USER %SELECT SLDS omits them).

    Contacts/Comments: To avoid these problems, don't switch logs after IMS start and prior to DB activity; have long RECOVPD/GENMAX; don't have long selector backlogs.

     

  • Name: Sheila Marsh – Bank of America – 0181-760-6504

    Problem: Has anyone stopped using BTAM or removed all BTAM definitions?

    Contacts/Comments: Shouldn't need them any more – talk to Dougie.

  • Name: Kevin Porter - Lloyds TSB Group – 0171-232-3819

    Problem: Year 2000 project has found program called IIBATCH = Inquiry IMS, where does it come from?

    Contacts/Comments: Think it is some old product that got renamed.

  • Name: Dave Grenfell – Integris – 01932-781000

    Problem: MVS 3.1.3 moving to OS/390, small DB2 and big IMS-DB. Can V4 run on OS/390?

    Contacts/Comments: Yes, but only Year 2000 compliant with PRPQ (see above).

IMS Quiz – Alan Cooper

Alan hit us with a series of IMS questions to tax our brains and get us ready for lunch.

1. What is the theoretical maximum size of a non-partitioned VSAM HDAM DB?
2. What PROCLIM value is equivalent to infinity?
3. When, if ever, does a HIDAM DB use RAPs?
4. What is the maximum number of programs involved in a deadlock that can be detected by the PI internal lock manager?
5. What is the ETO default Logon Descriptor name for a non-SNA locally attached printer?
6. What are the two commit modes with OTMA?
7. Which system checkpoint is used for an emergency restart?
8. Which RECON record is used to record the creation of an RLDS?
9. In n-way sharing, what is the purpose of the IMS DB Managers' use of the Coupling Facility?
10. Other than Transaction Manager, what else is IMS-TM?
11. What mode is used with APPC/IMS when IMS queues the input data?
12. Which IMS command is used to increase the maximum number of parallel copies of a transaction that are allowed?
13. What is the purpose of the AOIP?
14. What is the maximum number of VSAM sub-pools that could be defined in a DB/DC system?
15. What is the maximum number of message queue datasets?
16. What resolves a deadlock between IMS DB and DB2?
17. What is the last field of the IOPCB (NB IMS V5)?
18. On a PCB macro, what is the maximum possible value of KEYLEN?
19. Which utility uses USIDs?
20. At an IMS AUTO restart, how does IMS determine if it is a normal or emergency restart?

Answers are given below – see how well you do without cheating! Best performance at the meeting was six correct answers – well done Andrew Jack.

LUNCH

IMS Version 6, It's the Real Thing – Tom Ramey (IBM ISC), Dougie Lawson (IBM UK)

• We were hoping to get a real live German customer to talk to us, but they were too busy throwing away all their imported beef on the bone. So we got the IBM guys who are involved in all the early sites to come and tell us all about it.

• Tom gave a presentation covering the most important features in V6, with Dougie giving real live demos of various parts – clever stuff. Dougie even demonstrated entering a transaction on one IMS and getting the response on another one! This has serious implications if you are using MSDBs which aren't shared; hence the recommendation to migrate to VSO DEDBs.

• Note that Shared Message Queues in the CF are not mandatory – the old way of message queuing is still there.

• Don't have to recompile your programs when going to V6, despite changes in the IOPCB, because the change (addition of timestamp to the end) is not in your storage/program anyway – it is just a mask.

• A copy of the foils is enclosed.

AOB

• PQ03038 - V6 & V5 together – the rules on changing the clock back when the RECONs are shared between V5 and V6 (who would want to do that? Let someone else like Dougie or me or a customer test it out first and let you know how it all works!).

• II10131 – Upgrade procedure for RECONs when going to V6. Beware that the V6 migration and coexistence SPE required a different procedure from normal, because the key length of the RECON records changes.

• DSSN fixes – PQ05554 and PQ1554.

• If you have Omegamon, ask for QI28510 – problem is an 0C4 followed by a U0113 in the control region.

• DBCTL – due to low turnout at last meeting, looking at doing combined meeting next time.

• Please volunteer to present (or volunteer your colleagues).

Quiz answers

1. 40 GB = 10 DSgroups @4G each
2. 65535
3. With PTF root pointers, but no PTB
4. 64
5. DFS327P
6. 0 and 1
7. Last but one
8. PRILOG
9. Buffer Invalidation
10. Transport Manager (RSR)
11. Implicit
12. /CHANGE
13. To queue messages (from type 2 AOI Exit to type 2 AOI programs)
14. 16 pools, each of 22 subpools = 352
15. 10 x SMSG + 10 x LMSG + 1 x QBLKS = 21
16. SQL call timeout
17. GROUP-ID
18. 15 x 255 = 3825
19. Recovery
20. RDS header record