For a while, I’ve noticed that most IBM i sites that have Hardware Management Consoles (HMCs) don’t seem to show them much love. In fact, in many cases I would go as far to say that they actively go out of their way to ignore them.
If you have an HMC, ask yourself one question. When was the last time you backed it up? If the answer is less than a year ago, I want you to feel proud, as in my experience that already places you in the top 10% of HMC-lovers.
If you really want to go to the top of the class, ask yourself when was the last time you applied an update to your HMC? If you have done this at any point at all, then I already think you are an HMC superstar. But if you’ve done it in the last year, then, on behalf of your HMC, I want to reach out and give you a hug.
OK, I understand that loving your HMC is not likely to get you a pay rise and it certainly doesn’t make you more popular at parties (trust me, I speak from experience here – the ladies truly don’t appreciate a man who knows how to look after his console). But in times of need, your HMC is your best friend. And, as such, it deserves a little respect.
At this point, you may well be asking: what is he talking about, “best friend”? What’s so good about having a HMC? So let me share with you five reasons why I think you should love your HMC.
Reason 1 – It make your systems run faster
I’ve lost count of how many sites I have been to that have not tuned the performance of their hardware. Let’s just say that most people, if they ever do tune it, only ever do it once. This is, of course, crazy, as the way we use our systems evolves over time.
It’s true that you can do some amazing tuning within the operating system itself but if you have processor, memory or I/O that is being under-utilised by one LPAR on your system and there is another LPAR crying out for it, then shame on you.
There are loads of ways you can use your HMC to improve performance. Here are just a few:
• Using uncapped shared processors to dynamically allocate processor resource where it’s needed, when it’s needed while still guaranteeing minimum performance levels.
• Using active memory sharing to do the same sort of dynamic allocation described above but for memory.
• Using DLPAR to move memory and/or I/O between resources, either manually on a schedule or, if you are really clever, on demand.
• Enabling system-wide performance data collection so that you get a truly holistic view of where your resources are being used and where they are really needed.
• Coupled with Active Energy Manager, you can overclock or slow your processors to deliver what your system really needs.
• Coupled with Live Partition Mobility, you can move workloads between hardware hosts to enable access to extra resources without any downtime on the LPAR.
Reason 2 – Shared remote console
With the HMC, you get reliable access to your console both locally and remotely. This is fantastically useful, especially if you want to run an entire system save (ie Option 21 from the GO SAVE menu) remotely.
I highlight the HMC’s reliability for those of you who have suffered at the hands of IBM’s Operation Console in the past. I want to stress that you will not experience any of these reliability issues if you switch to HMC-managed systems.
But, reliability apart, the HMC also allows you to set up a shared console. This does exactly what it says on the tin; it gives two or more people access to the same console at the same time. This is great for support or for those of us who work in teams where you want to allow more than one person access to the console for sys admin tasks and don’t want to have to drop and reconnect a console session between tasks.
Creating these shared consoles is simple. You simply select the LPAR that you want a console for and select the “Connect Shared” option (see below).
The first person in the chain is asked to enter a session key. You simply tell all the other guys you wish to share with what this key is and they will be prompted for this extra password to gain access to the console.
This extra failsafe is designed to maintain privacy and system control – even between competing SysAdmins…
Tip: The only odd things about shared consoles compared to normal screens is that you don’t see what the other guys are typing and only the person who presses the enter or function key sees the Input Inhibited indicator (the X at the bottom of the screen) that signals the session is processing the request.
Reason 3 – Remote power on
Remote power on is especially useful if you have multiple LPARs. But even if you just have one LPAR running IBM i on your hardware, this gives you the ability to remotely reach in and power on both the system hardware and your IBM i LPAR (yes, the latter does not necessarily have to start when the hardware does).
Imagine you’ve had that power cut in the office (no pun intended) and your server is down. Typically, this has meant a trip into work simply to press the button on the front of the server to start it back up.
This is not the case when you have a system managed by HMC. You simply select the system and issue the “Power On” command (see screen shot below).
Reason 4 – Centralised email alerting
We all like to know when something goes wrong, particularly if it is out of hours. Before, if you had multiple IBM i systems, you had to write a series of monitors, one for each LPAR to alert you of problems occurring on them. This is complex and it does not scale well. Worst of all, if the problem is catastrophic there are no alerts as the system is down.
The HMC has a built-in, easy to use email alerting system that can notify you of major or minor events. It would be wonderful if there was an API to add your own alerts but it’s free, so let’s not complain too much.
In the screen shot above, we can see that emails are sent to our mail server Trigger. In this example, only major “call home” events are emailed to user Rodney whereas all events are sent to user Delboy.
Reason 5 – Reduce your system downtime
We have been used to swapping disks without downtime for many years but if you have a HMC you can, in most cases, swap PCI adapter cards, system processor nodes and even FSP firmware without downtime on your system.
These functions still amaze me even to this day. The technical issues IBM had to overcome to make this possible are truly legion and very complex. Sadly, for this reason, if you really want me to explain why, you will need two extra things. Firstly, about two hours of free time and, secondly, a bar from which to buy me several drinks while I explain it to you. The latter is to stop both our heads spinning.
What makes this process even better is that the HMC will guide you through the hot upgrade process step by step and with helpful explanations, diagrams and safety checks to make sure you don’t go wrong.
This information is displayed in real time. It is detailed too, with many screens displayed for each process. Plus it will match the model of the specific system you are working on.
Nice to see you…
Our next i-UG meeting will take place in central London, UK on the May 19, 2015. We’ve already confirmed a number of excellent guest speakers.
David Spurway will demonstrate how buying a Power 8 can be cheaper than keeping your current server. Andrew Ireland will be showing off the latest and greatest from Web Query.
Kevin Askew will do a show-and-tell on 3D printing and last, and by all means least, I will be talking about performance monitoring.
Hope to see you there. More details at the i-UG website . In the meantime, if you have something in particular you’d like me to discuss here on PowerWire, you can contact me via the feedback below or through my website.
Steve Bradshaw is the founder and managing director of Wolverhampton, UK-based Power Systems specialist Rowton IT Solutions and technical director of British IBM i user group i-UG. He has been a key contributor to PowerWire since 2012 and he also sits on the Common Europe Advisory Council (CEAC) which helps IBM shape the future of IBM i.