Evangelizing Mainframe
Print Email

Getting Linux Under Control

One of the consequences of running Linux on a mainframe is that you need to manage it. And, as we all know, you can’t manage anything without first being able to monitor it. So, how can you do this easily when it doesn’t produce any system management facility (SMF) records? Or, more particularly, how can you easily monitor what’s going on with Linux?

Linux on System z has been knocking around for 11 years now, but it’s only become really popular in the past year or so. If you search the Internet, you can find quite a few success stories in which people have transitioned from a large Linux server farm and a mainframe, to just a mainframe that’s also running the Linux images. There are basically two common ways to run Linux on a mainframe: You can run each image in its own LPAR, or run lots of zLinux images in the same LPAR under z/VM.

For the sake of completeness, there’s a third way, which is to run the Linux image on the whole mainframe. I haven’t found anyone doing that, though—but you can let me know if you know of anyone.

When looking to monitor Linux from a z/OS partition, there are several capabilities to look for. For instance, is it able to send monitoring information to a browser? This enables you to view information on a phone or tablet, as well as a laptop, etc. From a browser, you can see what’s happening on the Linux parts of the mainframe from almost anywhere in the world.

Several vendors enable this type of remote monitoring by running software on the z/OS partition and also a piece of agent software on each of the Linux images. Some agent software solutions are able to send Linux commands to monitor what’s going on, and send the results to the z/OS partition. It then formats the results and makes them available from a browser. With this kind of setup, there’s no need to configure the agent software because it’s the same for each Linux image.

Another important capability is a drill-down facility to get more information about what’s going on, and, perhaps more importantly, what’s going wrong. Some solutions allow you to click on a part of the display to open a new window for more information, or to mouse-over an individual Linux system’s icon to see what the Linux image is called, etc.

And if you can monitor Linux images using a standard agent, why not use the same technology to monitor Linux systems running on Intel boxes? Such solutions are particularly useful if they can monitor Linux systems running under VMware. In fact, so long as the Linux system has an IP address and the agent software has been downloaded, the monitoring software might even monitor that Linux system from your browser via the mainframe-based software. Now that would be brilliant.

In addition, having an extensible solution is beneficial. Some products allow the Linux systems—whether they’re on the mainframe or not—to be automated from the mainframe’s z/OS partition, using a familiar programming language, like REXX.

And finally, the best solutions offer multiple means of delivering the alerts when thresholds have been reached. Sometimes, people miss a patch of red on their browser. So you need to configure the software to send an e-mail alert to whomever’s on duty, or deliver a message on the operator console, in NetView, or even on the Tivoli Enterprise Console.

Consider these capabilities when examining software options to get your Linux environment under control. These features will enable you to better monitor and manage all your Linux images, wherever they are.

Trevor Eddolls is CEO at iTech-Ed Ltd., an IT consultancy. For many years, he was the editorial director for Xephon’s Update publications and is now contributing editor to the Arcati Mainframe Yearbook. Eddolls has written three specialist IT books, and has had numerous technical articles published. He currently chairs the Virtual IMS and Virtual CICS user groups.

Posted: 12/20/2011 11:40:37 AM by Trevor Eddolls

Print Email

Join Now!