Edit: My first post for IT Jungle
Originally posted November 30, 2020
There are philosophical differences that exist between people regarding the best ways to configure, manage, and maintain infrastructure, and this holds as true for infrastructure built around IBM i as it is for other kinds of platforms. The easiest and fastest thing to do – and what makes the most sense according to plenty of people – is to do what they have always done. Just keep replicating the past out into the future to infinity.
Others point out that this mindset stifles progress and ignores all of the innovations that have come to the IBM i platform through the years. How can you even consider not making the full use of your computer’s capabilities? Why are you still running your system the way you did in 2010, or in 2000 or even 1990?
A recent tweet by Larry Bolhuis sums up this way of thinking nicely, if a little hyperbolically:
“When you update your @IBMPowerSystems are you truly updating or just replacing one generation of CPU w/ the next and adding a dusting more disk and memory? If your business partner didn’t discuss PowerVM, SAN Storage, and VTLs before you bought, get a new partner.”
To expand on that, why would you ignore the power and flexibility to be gained by connecting to a SAN? Among many other things, SAN-based replication, snapshots, the ability to perform Live Partition Mobility, are given up when you refuse to move forward and think differently. Why purchase internal disk or an SSD that’s only available to one system in your environment when you could share in a pool of disk that’s used by other machines?
IBM resellers and business partners should strive to educate customers, and increase their own awareness. They should point out the advantages of, for instance, connecting to a SAN. Of course to assume that all IBM i installations even have SANs would be a mistake. Sure, they are present in most large environments, but they may still be the exception in smaller shops. One size does not fit all, and real-world circumstances must be recognized and acknowledged. Still, sellers should be sharing this information with you. Small shops may still benefit from connecting to a smaller IBM SAN, for instance. What’s important is that these discussions occur and customers learn of new options.
It’s a fine line, of course. Some choose to stay put. They’re satisfied, in many cases thrilled, with what they have. The system just runs. There have been no outages, and what more can you ask for from an investment in hardware? Everything is simple. Troubleshooting is straightforward; diagnoses are easy. The focus is the machine and its components, and possibly the network connection. Personally, whether it’s in my own environment or through my consulting with clients, I always get a charge out of working with new technology and getting the most functionality out of it that I can. But I can see the elegance in having a reliable standalone system, and in that way of thinking.
When it’s time for a hardware refresh, the request in this scenario is simple, something along the lines of: “Please give me the current equivalent of what I’m currently running.” Maybe a little more disk is needed, or a faster, higher-capacity tape drive. With the speed of today’s processors even a factory de-configuration of some cores is a consideration; if the workload in question is small enough, maximum horsepower may not be needed. So the order is placed and the upgrade is performed. You are content, knowing that this new configuration will last for years, or at least until this hardware reaches end of life and you’re compelled to do another upgrade.
I was recently part of an interesting back and forth with Bolhuis and others about these sorts of scenarios and choices. It was noted that customers are not always educated about what they can do with their Power hardware. They may not fully appreciate the benefits of virtualization, or running things differently than they always have:
“You know your @IBMPowerSystems can run more than just #IBMi or just #AIX! Use your next upgrade to enable deployment of both, or Linux, or all three. Don’t let yourself be boxed in by old single system thinking. Be sure your BP knows the options for virtualization!”
Especially for those of us who think about this stuff for a living, it’s natural to assume that everyone else is just as engaged. But not everyone attends virtual briefings or consumes every detail of IBM’s announcements, they have better things to worry about. (That is why you have The Four Hundred doing this for you, in fact.)
Replacing existing hardware with similar, more current hardware may make perfect sense, but remember it isn’t the only way. Business partners should be partners. They need to really listen to you, they should strive to understand your needs, answer questions and keep you informed. They should help you solve your business problems. And people like me aren’t serving our clients and customers if we don’t suggest alternatives. With some slight modifications, a traditional standalone system can host multiple LPARs, including Linux on Power workloads, and production and development LPARs, for example.
A position we do not want to be in is to get a request from management or from system users that we cannot fulfill. They may come to us and say that they understand that in theory they can add an LPAR to an existing Power System. Or they may tell you that they are under the impression that they can boot the system from a SAN. They may approach you and tell you that they understand that they can do X, Y, or Z.
The answer may very well be: “Yes you can, in theory, but since we did not discuss this, you’re missing an HMC, and you’re missing these other components that will enable that functionality. So until we remedy the situation, the answer is going to be no.”
None of this is a revelation; these capabilities have existed for years. For you the decision may not lie in knowing that you can do it, but in the incremental costs involved, or the skills available, or the thinking that if it isn’t broke don’t fix it, or as stated before the elegance of the system running as it is.
Some may view IBM i as a legacy environment and seek an excuse to discard it for something new, shiny and sexy. Education needs to occur at the management and executive levels as well. Power should be a strategic platform to your business, and wherever and however possible, it should be exploited to its full potential.
I think we can agree that some customers, and maybe even some sellers in the IBM i world aren’t up to speed — or simply not all that interested in learning about everything that the platform can do. That’s unfortunate, but that’s part of my takeaway from the discussion with Bolhuis.
Blame can go in all directions on this. Should IBM i operators and admins make more of an effort to seek out the information? Should IBM make information easier to find? Should employers be providing their employees with the training and classes and conferences they could benefit from? Should sellers be more active in educating themselves, and having conversations and whiteboard sessions and lunch and learns with customers?
It could simply be a product of the reality we all live in. Everyone is doing more, everyone is busy. Day to day it is easy to let learning take a back seat to current system issues. And that doesn’t even begin to get into the unprecedented chaos that has been 2020. With our day to day jobs, with our lives, it’s tough to balance everything. It all takes time, and there are only so many hours in a day.
Ultimately though, a lack of education or knowledge or experience can be remedied. The discussions are good – they’re imperative, really. Even internal discussions can be beneficial. Maybe ask yourself: How would life change if you went from a single production LPAR to having access to multiple instances of IBM i? Yes, there are costs to consider: licensing, additional cores, additional memory, additional disk, etc. Some of you will look at the costs and dismiss this proposition out of hand. But sometimes these decisions are based more on valuing the way things have always been done and fearing the unknown.
Getting a test machine is a good start when it comes to trying new things. No one wants to mess with production, so a test box or partition is the best way to try something new or get better at what you do. Implementing a dedicated environment where you can change code and test backups, upgrades, installs, and so forth and never have to worry about affecting the business is really a no-brainer decision.
So let’s initiate these conversations. What would make your life easier? Where would you benefit from taking a slightly different approach? If you could change anything, what would you do differently? Needing to make no changes at all is also an acceptable answer, provided you have done your due diligence and spent a little bit of time considering all of your available options.
Rob McNelly is a senior Power Systems solutions architect doing pre-sales and post-sales support for Meridian IT, headquartered in Deerfield, Illinois. McNelly was a technical editor for IBM Systems Magazine, and a former administrator within IBM’s Integrated Technology Delivery and Server Operations division. Prior to working for IBM, McNelly was an OS/400 and IBM i operator for many years for multiple companies. McNelly was named an IBM Champion for Power Systems in 2011, an IBM Champion Lifetime Achievement recipient in 2019, and can be reached at rob.mcnelly@gmail.com.