Monday, November 29, 2010

Information on IBM Support

Here is a link to an interesting web page that provides a lot of information on using IBM Support and items like the Support Portal and My Notification. There are links on this page for instructional videos and a software support handbook. This is a good starting point for navigating the various facilities of IBM support.

Address space CPU usage info in OMEGAMON XE For z/OS

APAR OA34505 has some interesting information about the z/OS Address Space CPU usage display. Apparently it was noted that there was an inconsistency between how CPU % usage was being calculated in the OMEGAMON XE Tivoli Portal Address Space workspace versus how the comparable data was being calculated and displayed in the OMEGAMON II 3270 interface.

It seems the calculation in the TEP has been the sum of SRB CPU + TCB CPU. However in the OMEGAMON II for MVS interface the calculation has been SRB CPU + TCB CPU + Enclave CPU. The net result is that when looking at tasks in the TEP, you may see lower than expected CPU usage numbers.

For more information on the APAR, here's the link:

Wednesday, November 24, 2010

Adding HTTP information to your end to end view

In earlier posts I've written about how you can use OMEGAMON XE For Mainframe Networks to add network level information to your mainframe monitoring views. For example, you can monitor network traffic to/from CICS, IMS, or DB2, and look at such things as byte counts, round trip time, round trip variance, and network excpetions. This type of information can be a useful supplement to host response time captured by OMEGAMON XE for CICS or IMS.

Another interesting set of metrics you can add to the mix is HTTP monitoring. Many applications in most enterprises have a flow that may start with some form of user interaction at an HTTP server. That HTTP server will often interfface with some form of middleware (MQ, WebSphere, J2EE, you name it), and then to either CICS, IMS, DB2 on z/OS or all of the above. To have a more complete monitoring picture, it's nice to have HTTP information as part of the monitoring view.

Depending upon which Tivoli monitoring components you have licensed, you may have access to a tool called the Universal Agent (UA). UA is an interesting tool in that you can use it to pull in information from a wide variety of data sources. Among the data providers is file, socket, SNMP, and HTTP. The HTTP data provider allows you to monitor URLs, and track the status and response time of these URLs on an ongoing basis in the Tivoli Portal. UA is a great tool to add the additional layer of performance/availability data to the Portal.

Here's how you can add the information to your portal view. First you need to enable the HTTP data provider as an option in UA. To do that you go to Manage Tivoli Services and select the option to configure or reconfigure UA. When you click through that dialog you will get to a notepad editor pop-up. There you add HTTP to the start-up options (as I show in the example). You save the file, and re-cycle the UA process. The next step is to add a URL to be monitored. From the navigation tree in the portal, you select a Take Action, and select Add URL (again as I show in the example). You enter a URL to monitor in the pop-up, click OK and you are ready to go.

The data you get is interesting, and useful. For each monitored URL you will get a status, and a response time, along with page size and other information. Good stuff to add your portal views.

Monday, November 15, 2010

Tivoli User Community webpage

A good source of information on Tivoli products, plus information on such things as Tivoli user groups and other events, is the Tivoli User community web page.

Here is a link to the Tivoli User Community web page:

Thursday, November 11, 2010

IBM Tivoli is looking for your feedback

IBM Tivoli is piloting a new request for enhancement (RFE) web site for a subset of the IBM Tivoli products. The product list includes TADDM, ITCAM for Transactions, OMEGAMON XE for CICS, OMEGAMON XE for IMS, OMEGAMON XE for Storage, OMEGAMON XE for z/VM and Linux, OMEGAMON XE on z/OS, ITRPM, TSPM, and TWS.

Customer input is important to improving the products, and I encourage everyone to check out this web site, and add your input. I can speak from experience, that I can talk to product development and R&D all day long, but customer opinion is what counts most.

Here's the URL:

Wednesday, November 10, 2010

Improvements to OMEGAMON DB2 Near Term History

OMEGAMON XE For DB2 has a Near Term History (NTH) function that is one of the more popular and useful features of the tool. In NTH you can go back in time and look at detailed historical information at the DB2 subsystem (Statistics) and/or the DB2 application (Accounting) level. This is essential information for being able to analyze performance issues after the fact.

DB2 Accounting records, in particular, have the potential to consume a fair amount of DASD space. The records themselves are large, and DB2 may generate many of them, as many as millions a day in many shops. How far back in time you can go in NTH is a function of how many records you need to store, and how much space you allocate in the NTH collection files.

One little known aspect of NTH is that you have the ability to allocate and use more than the three collection datasets you get by default. It used to be you could go up to ten datasets. Now with recent enhancement, you can go up to twenty history collection datasets. This allows for even more storage space, and the ability to keep more history data online.

For more information on this enhancement, check out this link:

Wednesday, November 3, 2010

SA System Automation integration with the Tivoli Portal

The Tivoli Enterprise Portal (TEP) is more than just an OMEGAMON thing. Many other Tivoli tools, such as the storage tools and SA System Automation also integrate within the TEP. I've been working with a customer on an availability dashboard concept for managing z/OS. One of the tools at our disposal is System Automation, and we have it integrated with the TEP.

In the example I show, we have some of the information from the SA agent, and I use that data to alert on key z/OS task availability. One aspect of task availability is to alert not only on when a task is down, but to factor in that there may be times when the task is supposed to be down (such as for maintenance, etc). The SA agent interface provides information, such as resource status, but also shows the desired status for the resource. In other words, you can alert on when the task is down, and also factor in if the task is supposed to be down.

In the example, I show part of the prototype dashboard, and I show an example situation. Here I'm looking at both the status of the resource, and the desired status. In other words, only alert if the resource is down, but it is supposed to be up (at least as far as SA knows).