Posted on by Roy Boxwell

2026-05 Central Canada Db2 Users Group 2026

I love this User Group – Great people, great location at the IBM Lab in Toronto and , this year, a great lunch! (You had to be there last year to understand this…)

This year they went all out and we got a full “Canadian Database Day” on the Monday split into a Db2 track called “Db2 for z/OS, Everywhere, All at Once!” – which covered, for Db2 z/OS, Bufferpools, Application dev, DDF Applications and Locking, Logging and Security. For Db2 LUW you got Semantic search, AI Systems and Amazon RDS. Then in the parallel IMS track everything about IMS as this old bird is still firmly nailed to its perch and rumors of its death have been greatly exaggerated!

And then?

Then we kicked off on Tuesday and Wednesday into the normal grid style with four parallel tracks. Now, as always, I must state that I did not go into *every* session so what I report here might just be a cut-and-paste from the grid and changes might well have happened that I did not know about! Further I did not go to any IMS or LUW tracks even though I started my Mainframe career writing IMS accessing COBOL programs with IMS 1.3 – Yep – I am THAT old!

Db2 Z DBA Tracks

Db2 Z Misc/App-Dev Tracks

Unlike the IDUG all the above links *should* take you directly to the download file. If not please contact the CCDB2 group and not my good self.

The missing link?

As of the 26th May the following presentations are *not* at the link address, just the abstract and speaker details:

  • ZOS 02 and 03
  • ZMISC 01, 02, 08 and 09

I hope that these presenters get their slides off to the CCDB2 organizers as soon as possible.

Reviews

ZOS-01 Akiko gave us her usual excellent stuff all about what we have and where we are going. Interestingly she mentioned that with Db2 13 FL509 we must run the REORGs for directory and catalog migration to UTS. This is *not* done by a CATMAINT. A nice tidbit was that the two new RTS columns NSYNCREADIO are not updated for Utilities which makes them very useful for utility decisions! She also reiterated that the CDDS functionality is now included in the base Db2 install and so we can all use this feature to get easy reachable copies of our compression dictionaries! Check out my blog all about using this facility called “2025-01 Compress and Replicate This!” at this link www.segus.com

She then went into what she called “Part 2” and mentioned that the new DDL Clauses FOR DGTT or FOR SORT got two APARs for problems: PH65632 and PH70190

Also note here that FOR SORT should use more than one for the MAXPARTITIONS value to benefit from the above performance improvement fixes.

Looking a little bit into the future she listed a few Scalar functions which are no longer deemed “Quantum safe” meaning that perhaps not today but certainly soon this level of encryption will be easily broken by a quantum computer. ENCRYPT_TDES or ENCRYPT for example.

Then she mentioned they are looking at changing the Lock Structure to get multiple versions which would make Datasharing even more scalable and resilient! Finally, she touched on a new possibility to access index’s using different column orders. No need for one index on COL1, COL2 & COL3 and another index on COL3, COL2 and COL1. This would be incredibly cool if they pull it off!

ZOS-02 Mark and Tori showing us all the fantastic things you can now do with System Profile Monitoring. If you have not looked into this start now! Just remember that START PROFILE starts ’em all and STOP PROFILE stops ’em all! This might *not* be what you wanted or expected! It was also repeated that user *must* start using Client Info fields. This should be standard but some shops still have one technical user id for *all* their CICS transactions so it is *impossible* to filter correctly or usefully. The other bug bear that is always asked is “Why cannot I filter on User Id and Location?” The answer is when it connects it knows just the location then a long time later it finally gets the User Id…Basically the code paths cannot handle this. If you are using cloud the Location address is normally very bad for filtering and so you must get Client Info added! A few interesting bits is that most occurrences of MAXDBATs being hit apparently are “ghost” transactions that just connect and sometimes run an SQL against SYSIBM.SYSDUMMY1. Then sit there doing nothing. Purely to see if Db2 for z/OS is there or not. Db2 for z/OS is always there, so if you can turn off these “ping” SQLs it would be a very good idea! Last bit was that the number of active profiles is limited to 4096… Do not exceed this number!

ZMISC-03 Ulf showed a relatively simple way to completely, safely and with no risk, copy your entire production metadata down to a sandbox and still guarantee access paths as if from production. Lots of people forget silly things like RTS or number of CPs & BUFFERPOOL numbers and sizes. Check it out to see how it can help you pre-check your next set of Db2 for z/OS APARs, Db2 Version release or machine upgrade or even downgrade if consolidating.

ZOS-04 Richard showed me how old I was as my data strategy is very very old indeed! More buzzwords than you can hit with large stick! The brave new would of data everywhere all over and only replicated when it absolutely must be. Great stuff! Favorite statistic “25% of all CPU MIPS is spent replicating data”. Discussion topic – Latency: 0 if virtualized to “it depends”. It depends on z/OS load, Network load and Target system, normally LUW, Load.

ZOS-05 John doing his usual fantastic stuff this timing taking aim at High Performance DBATs. One Golden Rule: Only for high volume simple transactions. The problem is to find them in all of your workload! They are a trade off : More memory & More DBATs versus Lower CPU. Tracing is good with IFCIDs 411, 412 and the best is 365. The one disadvantage is that at least one of the packages they use MUST be bound with RELEASE(DEALLOCATE).

              He then mentioned a small change in behavior that crept in with the NULLID packages in that they all got changed to RELEASE(DEALLOCATE) and he strongly advises everyone to REBIND them all with RELEASE(COMMIT) and, if you want High Performance DBATs, create a new COLLID where the packages are there bound with RELEASE(DEALLOCATE) and use System Monitoring to redirect to this new COLLID or change the JDBC path.

              When you issue the -MODIFY DDF PKGREL(COMMIT) this is not an *instant* fix to switch these things off. It will take time for all accessing threads to eventually be reset. Remember the DBAT only gets recycled after 500 times in Db2 13, before it was after 200.

ZMISC-06 Michael walked us through a great review of things over the years that come and sometimes go! That people tend to neglect or forget. ZPARM DISABLE_EDMRTS for example is very good one! RUNSTATS not updating the RTS in certain scenarios is another. A quick review of all the new and snazzy commands we have now got including DISPLAY BLOCKERs of course. My personal favorite due to it being the *only* command that outputs the Timestamp when it was executed. Sounds minor but if all you have is command outputs and you have to work out who did what when and where then timestamps are great. The details and hints about use of RECOVER with BACKOUT was also very illuminating!

ZOS-07 Roy doing his usual great stuff (What else am I going to say? 😊 ) and wrestling with the darn mike as it did not want to work with the British guy! Nice questions about how long til doomsday etc etc. Just download and run our freeware tool: MigrationReadiness HealthCheck for Db2 z/OS at www.segus.com to find out, very quickly, exactly where you currently stand. The worst things are SYNONYMs, Work tablespaces and TCP/IP. I have customers up in the 8,000 synonyms and some with zero. I guess it all goes back to when you started down the road of Db2! With work files if you have 400 on 87 subsystems that could be a lot of work. Here the DB2 13 FL508 DDL enhancement FOR WORK / FOR DGTT will help as well but you must still DROP those spaces!

ZMISC-08 Mark and Anna with standing room only! Absolutely packed room where no-one even offered me a chair – The Audacity!!! The dynamic duo gave us all their data about the lifecycle of packages for static and dynamic SQL diving down into all the grubby details that we all love so much! The question of the day was “How does Python on the Mainframe get allocated to a Package, and which one?” No one in the room knew the answer! The hope is directly to some package or possibly even using Type 4 going out and back in again on the network! I have noted this down as something to try and find out!

Update: Found out from Jørn Thyssen that it will use ODBC and then CAF or RRSAF with default PLAN DSNACLI and one of the seven packages within. The BINDs are all in <your Db2 HLQ>.SDSNSAMP(DSNTIJCL). Sadly I still have no idea which package but probably DSNCLIC1. Here is a great BLOG all about Python including, right at the end, a hint about APF Authorizing its load lib to get Python ZiiP enabled… Pretty neat! Python ibm_db on IBM z/OS UNIX (USS): a field ready install & troubleshooting guide

ZMISC-09 Kranthi showed us all the details and intricacies behind the “certificate” problems we all have or will have! With the lifecycle of certificates dropping very quickly you must try and get automated systems to cope with the renewal. For a few PCs or laptops and phones it is ok to do it yourself but in a firm? Fun fact: How many certificates does your average lap-top have installed? Between 100 and 600 is normal(!)

They think it’s all over!

Then we had to leave to get our flights but I was very happy indeed to be there again and I already look forward to next year and being there once again!

The Food?

Chris, the food was excellent this time but I still remember that Beef Bacon horror at the IFL!!!

Many many thanks to all the volunteers who put on the whole show and IBM for allowing the whole thing to be run from their Labs and I personally thank all the IBM Employees who were asked nicely to “leave the Canteen at 12:30” so a bunch of DBA crazies could sit down and have lunch!

Lest we forget!

For those who are now retiring and not there next year:

Thanks for all you have done and enjoy your curling!

TTFN,

Roy Boxwell