(Originally posted 2007-06-23.)

Here are my abstracts for the conference:

Session B11: DB2 Data Sharing Performance for Beginners

This presentation provides an introductory-level view of how to look at the DB2 Data Sharing performance numbersfrom both a z/OS / RMF and a DB2h perspective.

Performance topics include: XCF, Coupling Facility, Data Sharing Structures, The application’s perspective, and Structure Duplexing.

Performance topics don’t include: Other forms of Data Sharing eg VSAM RLS, and overly detailed descriptions.

Session P22: Memory Matters in 2008

For z/OS LPARs memory management has changed radically over the years – from both the operating system perspective and that of applications. And the pendulum has swung back and forth between focusing on Real Memory and on Virtual Memory.

This presentation discusses managing both Real and Virtual Memory – from the perspectives of both the operating system and the exploiting products. The products include DB2, DFSORT, CICS, IMS, MQ and Websphere. One topic of particular importance to installations upgrading z/OS is the Release 8 Real Storage Manager rewrite.

Session P23: Much Ado About CPU

zSeries and System z9 processors have in recent years introduced a number of capabilities of real value to mainframe customers. These capabilities have, however, required changes in the way we think about CPU management.

This presentation describes these capabilities and how to evolve your CPU management to take them into account. It is based on the author’s experience of evolving his reporting to support these changes.

(Originally posted 2007-06-21.)

This customer is using DB2 Hiperpools, despite being on a z9 BC processor and running DB2 Version 7.

What I notice from DB2 Statistics trace is that most of the Hiperpool pages are not backed by memory. In other words they’ve ceased to exist. This despite there being plenty of memory in the LPAR. One thing that’s also significant is that there are very few reads back from the Hiperpools. It’s mostly writes.

Somehow I think the two facts are linked. In any case I would recommend the customer moves to Version 8 and, in the short term, uses Dataspace pools instead of the combination of Hiperpools and Virtual Pools.

Meanwhile I’m still scratching my head. 🙂

(Originally posted 2007-06-20.)

… with a Performance Consultant? 🙂

Seriously, here’s something that made me go “hmm”…

I started looking at the latest set of data from a customer…

Their biggest-CPU WLM workload is called “BATCH” when they told me they were a CICS shop – and that CICS was their main CPU consumer during the day.

But it turns out that the biggest service class within that workload contains exclusively CICS regions. (And it is called “PRDONL”.)

So that’s all right then. But you can see why I’d be confused. 🙂

Moral: A name is just a label, not a description. At least that’s true with WLM. But funny how we read things into names that generally are there. 🙂

(Originally posted 2007-06-13.)

Regular readers would know I’m working on my “Memory Matters in 2008” presentation, which is a re-spin of the ’07 Version. One of the things I talk about is IMS.

Thanks to my team mate Andy Wilkinson for this list of IMS Version 10 enhancements, all of which are virtual storage usage improvements.

Here are the more important items:

• OTMA message flood protection (prevents LSQA filling). This is retrofitted to IMS V9 by PTF.
• New IMS log records (x’4511′ and x’4512′) to report on control region memory use.
• IMS V10 acquires CSA for Fast Path resources in discrete pieces, rather than insisting it all be contiguous, which helps large Fast path customers

And here are some other, less important, IMS V10 items:

• ACBGEN supports 31-bit addressing for the first time, which helps programs which access thousands of databases.
• PCBs are passed from CICS to IMS using 31-bit addresses (instead of 24-bit addresses), which also helps programs which access thousands of databases.
• Some IMS modules move above the 16Mb line, saving about 100KB of virtual storage.
• Several IMS modules that used to reside in 24-bit common storage now reside in 31-bit common storage

The details of these are in the IMS V10 Release Planning Guide GC18-9717-00 available from here.

(Originally posted 2007-06-06.)

One of the problems with having a presentation with a topical title is that you keep having to update the darned thing. 🙂

“Memory Matters in 2007” is a case in point. It’s morphed over the years from two presentations: “Central Storage Performance” and “Expanded Storage Performance” (the latter of which eventually became obsolete) to a presentation that covered OS/390 real memory, DFSORT and DB2 virtual storage, to the current “Memory Matters in 2007” which covers z/OS real memory and a bunch of subsystems (CICS, MQ, DB2, DFSORT, IMS and Websphere Application Server).

Now I’ve been asked twice to present the same topics but fast-forwarded to “Memory Matters in 2008” – when we’re still in 2007.

I suppose I could call it “Memory Will Matter in 2008″ but that rather spoils the pun. (I do wonder how many of my audience in München got the pun.) 🙂

So anyhow, I’m going to update most of the sections to cover eg DB2 Version 9 and CICS TS 3.2. Just a little sooner than I’d anticipated. 😦

The second outing for the ’08 model will be at the UK GSE Conference October 30-31. This is at Chesford Grange, near Leamington. (Haven’t I been here before?) 🙂

What I might try and do is post new foils (or at any rate sketches of them) here and see how people react.

And I suppose my manager will want me to podcast this. 🙂 I wonder where to host such a thing.

(Originally posted 2007-05-31.)

(Also posted to MXG-L Listserver (MXG-L@PEACH.EASE.LSOFT.COM) which I highly recommend as a place where mainframe performance people hang out.)

I must confess I feel slightly foolish about this… 🙂

After having asked for input into my UKCMG BOF one of the themes that came out was better ways of doing what IFASMFDP does. (I already have this on my agenda – as you’ll know if you follow my blog.)

So I asked my friend Frank Yaeger (who is the DFSORT Function Development lead) the following question:

“Frank, I thought you told me once that DFSORT had some issues with handling actually spanning records. Did I hear you right?”

to which I got a reply (paraphrased)…

“No, that’s wrong. DFSORT handles spanned records just fine. We can handle records up to 32767 bytes in length”.

So, a healthy slice of humble pie is mine, complete with plastic fork and paper plate. 🙂

Seriously, would anyone care to do a test for me and copy some eg Type 30s using DFSORT? 42-6 and 74-1 are the other cases that bother me. And perhaps 74-5. The test would be to copy and then run the results through (say) MXG. (I suspect using ICEGENER wouldn’t be a fair test as it may well revert to IEBGENER which we’ve always said does break SMF records.)

What I’d like to do, if such tests are successful, is to publish some examples of using DFSORT as a “better IFASMFDP”.

One thing still bothers me a little: Using DFSORT Copy doesn’t produce Dump Header and Trailer records – SMF Types 2 and 3. Does anyone actually care about losing them?

Note: This is emphatically not a DFSORT vs Syncsort question. I’d assume Syncsort also copes with spanned records. If someone is in a position to test that then great. If this were ever to be a DFSORT vs Syncsort thing it’d be on the grounds of what you can do with DFSORT vs Syncsort, not their ability to cope with SMF.

(Originally posted 2007-05-25.)

I happened to be travelling through Heathrow Terminal 2 yesterday and the day before. Two incidents that made me scratch my head, both related to Passport Control…

• On the way out there are two Passport Control desks, just past Security. They are set a few feet apart and people can go either side of either of them. For once there was someone on duty, but only one officer. At the head of the queue is a man who asks the Passport guy “Who are you? And can you prove you are a Passport official? You could be anyone.” The debate ensues for a few minutes, with the queue getting longer. Finally the official pulls out his badge and shows the man. The man’s claim is that there were no signs up announcing this as a passport checkpoint.

  Meanwhile some other passengers go for the gaps between the desks. Finally the passport official waves everyone through the gaps.

• On the way back in there are (as most people know) two queues: EU and non-EU. The EU queue is long and there are 4 people on duty at the head of it. There are more people on duty at the head of the non-existent non-EU queue. Nobody seems to think to let some of the EU travellers come over to the non-EU queue. Next time this happens I think I’ll try the non-EU queue and see what happens.

So this seems like chaos to me. Perhaps it’s just as well that Terminal 2 is being pulled down next year. On the front of the building is a nice sign saying so (tho’ I did already know this: A taxi driver told me last year so it must be true.) 🙂 Also on the sign is an artist’s impression of the new Terminal 2. It looks like a giant inflatable mattress without jetways. I wonder how you’ll be expected to board your plane: Bounce onto the mattress and bounce in through the open door? 🙂

(Originally posted 2007-05-18.)

As you probably know this year’s UKCMG conference in June 25-27, Whittlebury Hall, Northants.

I noticed when I read the agenda yesterday 2 things:

• Glenn Anderson’s being worked awfully hard. He’s a great speaker so do make the effort to attend his sessions.
• There’s a BOF session on each evening – just before the entertainment.

So I rang Laura at UKCMG up and offered to run a BOF. I got a “that’s highly irregular to book one beforehand 🙂 but a great idea” response from her.

So here’s what I’m proposing…

z/OS and Mainframe Performance Instrumentation BOF

As performance folk we quite often run into frustrations with the performance data, most notably SMF. I could articulate quite a few of those myself. Similarly I’ve found limitations in what the IFASMFDP SMF Dump Program does. Accordingly I’d like to gather a “wish list” of such items and discuss them with the customers who I know are at UKCMG. I’d hesitate to call them “requirements” as this isn’t a formal requirements-gat>hering session. (I have no formal remit for this – but I do have the ear of a number of key developers.)

And I’d obviously like to continue the dialogue – perhaps at other conferences, but certainly by conventional means.

On the Monday I will’ve talked about CPU and Memory instrumentation. That ought to get people thinking.

So please bring your ideas – and perhaps have a think about it beforehand. And feel free to contact me to “put stuff on the agenda”, though I have no formal agenda for this BOF.

And see you at Whittlebury Hall in June!

(Originally posted 2007-05-16.)

I’m reminded of this white paper on XCF Performance from Joan Kelley and Kathy Walsh. It’s a good read and has been recently updated.

It does also remind me that there are “D XCF” subcommands that give information that isn’t in the RMF data (SMF 74 Subtype 2). We process these records in our consulting and I’m getting considerably more familiar the underlying data.

(I also had a situation recently where large messages went through the coupling facility, whereas the standard messages had no opportunity to do so.)

(Originally posted 2007-05-12.)

Now I know why, as previously reported here, there’s a new power meter on the HMC. See this blog entry and the links from it for more details of the “Big Green” energy efficiency initiative.

And this goes way beyond the “packets of grass seed” campaign of last year – which highlighted the already available energy-saving characteristics of the mainframe.