Friday, February 26, 2010

Pulse 2010

I went to Pulse in Las Vegas this week. Pulse is IBM Tivoli's flag ship trade show and educational event for Tivoli technology. While there wasn't what I consider a huge amount of z specific content at the show, there were some good sessions on ITCAM and integration with OMEGAMON, response time monitoring, cloud computing, Linux on z, alert management, and other interesting topics.
If you didn't get a chance to go to Pulse, later this year there will be Pulse Comes To You, where IBM will take the event on the road. Here is a link to info on Pulse Comes To You. In the coming weeks check this link for more info:
http://www-01.ibm.com/software/tivoli/pulse/pulsecomestoyou/

OMEGAMON DB2 history collection options


I was visiting a customer recently and the topic of collecting historical data using OMEGAMON DB2 came up. I made the statement that you probably have as many as a dozen different ways to collect history data using OMEGAMON DB2. Couple that with the fact that DB2 has the potential to generate large volumes of trace and history data, and it points towards the need to consider your options when setting up history collection with OMEGAMON DB2.

To prove my point I went through this chart with my customer. You can have DB2 itself send data to SMF, and then run history reports from the SMF data. You can collect data to Near Term History (NTH). The primary use of NTH is for near term viewing within the Classic interface, but if you use the VSAM SEQ option you can also take the data from NTH and send it to other history processes. You can collect snapshot history, which can snap thread and subsystem data, and view that snapahot data within the PE GUI interface (shown at the bottom of the chart).

Which then brings us to Performance Database and Performance Warehouse. It may not be immediately obvious, but there is a difference. Both are DB2 trace data in DB2 tables, but with Performance DB you create and manage the objects yourself from hilev.RKO2SAMP members. Performance Warehouse is automatically managed by OMEGAMON DB2. Other than that, they are pretty similar. My preference is to manage my own objects, because of the size of the data involved.

But we are not done. There is also history in the Tivoli Data Warehouse. This is snap shot data, good for trending analysis, and viewable in the TEP, and using Tivoli Common Reporter. Plus we have many ways to collect data in the form of Collect Report Data (CRD). You can use CRD via the PWH, via ISPF interface, or via batch jobs provided in RKO2SAMP.

Quite a list, huh? I will probably be doing an article on this for Tivoli z Advisor.

Friday, February 19, 2010

ITM 6.22 changes history collection options


There are quite a few changes in the layout and appearance of the TEP with ITM 6.22. New icons, new screen layouts, a new tool bar layout, and new history collection setup is part of ITM 6.22. The screen shots I show here show an example of how the history setup is different with ITM 6.22. If you look at the example, you start with panel on the upper left, where you can select the type of agent for which you want to collect history. In the example I've selected z/OS. Underneath the z/OS selection you see several selections for which history collections have been enabled. When you click on one of the defined collection options, you see the other screen on the lower right. Here I've clicked on the distribution tab, and this is where you can specify what managed systems you want to collect history for.

With ITM 6.22, you have more history options, more flexibility, and more ways to control what history data you wish to gather, and how the data should be warehoused.

Wednesday, February 17, 2010

Some important new ehancements

My colleague, Wayne Bucek, in his Wayne_z_World blog has some information on a recent set of enhancements to OMEGAMON and ITM that optimize queries and displays that may request large numbers of rows of information. The more data requested, the bigger the savings. Like in the example Wayne mentions, if you are monitoring large numbers of MQ Q managers, this can have a big savings.

Here is a link to Wayne's blog, and the specifics of the maintenance:
https://www-950.ibm.com/blogs/9051592f-7640-466e-8524-6ae7dcc20c79/entry/monitoring_large_numbers_of_queues_better_faster_cheaper17?lang=en

Monday, February 15, 2010

Planning for ITM 6.22

If you are working on a plan to get to ITM 6.22, here is a link to some good information on platform support and compatibility for ITM 6.22. This link will show in detail information on supported operating systems, platforms, databases, etc.

http://www-01.ibm.com/support/docview.wss?uid=swg27018000&myns=swgtiv&mynp=OCSSZ8F3&mync=R#Database%20-%20DB2_R-swgroup

View threads in a data sharing group from OMEGAMON classic


For those OMEGAMON DB2 users that run DB2 in a data sharing environment, this is an enhancement that was added a while back to the classic interface, but you may have not noticed. The top line of the classic interface is the command line. The left portion of the command line is where you can enter various commands, screen space names, and navigational commands. The data sharing group support added some options on the right portion of the command line. You can now specify the DB2 subsystem, and if you want a single system (S option) view, or the data sharing group view (G option).

In the example I show first, on the top left, a single system view (DB2S option S), and notice there are no threads to view. I then specify option G for the data sharing group view. Now I can see threads connected to the DB1S DB2 member in the data sharing group. Also, I can drill in for thread detail, just like any thread in the classic interface.

This is a handy navigation option that you may, or may not have noticed before.

Thursday, February 11, 2010

Situation overrides provide added flexibility


I've talked before about situation management, optimizing situation usage to make situations more relevant, and doing things like using policies to manage situations, such as for time of day processing. If you are running ITM 6.21 or above you have the ability to take advantage of an interesting function called Overrides. The override mechanism provides a way to specify a different threshold settings for different requirements, such as Prime shift or Non-prime shift.

I include an example of how to use overrides. To access the override logic you have to go to the distribution tab for the situation editor, select a managed system and then click the override button. You can then specify the value setting, and here I specify if the setting is for Prime or Non-Prime shift. It's an interesting option, and I plan to work with it some more to see how it may be effectively exploited.

Tuesday, February 9, 2010

DB2 10 for z/OS goes beta

Being a DB2 guy from a ways back, this is a cool announcment. DB2 10 For z/OS is going beta. Here is a link for more info.

http://www-01.ibm.com/software/data/db2/zos/db2-10/

Friday, February 5, 2010

OMEGAMON DB2 Near Term History


OMEGAMON DB2 has a very useful Near Term History (NTH) function. NTH provides an easy way to be able to retrieve and review DB2 Accounting and Statistics records from the past few hours of DB2 processing. The data is stored in a set of VSAM files allocated to the OMEGAMON collection task. How far back the history goes depends upon the size of the files and the amount of data being written to these files. Now some of the data volume is driven by the DB2 workload activity. Accounting records are typically written when a DB2 thread terminates processing, and it is the Accounting data that is often looked at by the analyst when studying what DB2 applications have been doing. Statistics records are created on a time interval basis. Usually, you will have much more accounting data than statistics data. Also, OMEGAMON has the ability to pull in additional trace IFCIDs to get information on things such as dynamic SQL activity.

To understand the amount of data being gathered by NTH, there are displays that show the number of records written to the NTH files, by type. In the example I show, you see an example of common NTH settings/options, and then you see the record count in the NTH record information display. If you look carefully you see that 'Perf-Dyn SQL' has a lot of records written relative to the other record types. This is a good way to understand the impact of enabling certain collection options, such as dynamic SQL collection, and see how many trace records are being gathered, as a result.

Thursday, February 4, 2010

Did you know? About ServerPac....

Tivoli products on z, like OMEGAMON , can be ordered from Shopz in the form of a ServerPac.
You get the benefits of preinstalled datasets, and current maintenance already in those datasets, and the ability to configure the tools more quickly. Among the products supported include TWS for z/OS, IBM SA for z/OS, Tivoli Decision Support for z/OS (various components) , NetView for z/OS, ITCAM for SOA, and of course OMEGAMON (DB2, CICS, z/OS, IMS, Storage, Mainframe Networks, Messaging).

If you are pondering rolling out V4.20, this is a good way to get started quickly.