Greg Hintermeister dives into what Power System resource management looks like inside Flex System Manager, demonstrating that FSM can be your one-stop appliance for resource management.
Recently, I introduced you to IBM Flex System Manager (FSM) and gave you a taste of what FSM can do for you in managing your IBM PureFlex systems. FSM is an appliance dedicated to managing the compute, storage, and network resources required to run your PureFlex systems. FSM helps manage the chassis, firmware, virtual servers, switches, and storage, and it even helps deploy images into your PowerVM (or Kernel-based Virtual Machine—KVM) virtualized compute nodes.
But did you know that FSM also acts as the Hardware Management Console (HMC) for your Power System compute nodes? This means that FSM can assist you with managing up to four chassis that are filled with Power System compute nodes, as well as the switches and storage that the compute nodes depend on. In addition, FSM can perform detailed PowerVM actions related to your virtual servers (partitions) and other hardware-focused actions on your Power System compute nodes.
In this article, I’d like to dive a little bit more deeply into what Power System resource management looks like inside FSM. By the end of the article, I’m hoping you’ll understand that FSM can be your one-stop appliance for managing everything you need to ensure that the resources that are running your workloads are performing in top shape.
First Look at FSM
When you log on to FSM, you’ll see your chassis list. Click one of your chassis to view the chassis map (shown in Figure 1).
Figure 1: Chassis map with two Power System and four x86 compute nodes
Because a Power System compute node is selected, I can see a list of common actions at the right and an Actions drop-down menu at the bottom of the screen. The common actions list provides quick access to several tasks, and the Actions drop-down shows all the Power System management actions you can perform. Figure 2 shows another FSM-managed chassis and its Actions menu.
Figure 2: Context menu of a Power System compute node with Hardware Management Actions
Hardware and PowerVM actions are located in Hardware Information, Operations. The System Configuration menu option is used for actions such as creating a virtual server (partition), and Service and Support is used to perform a variety of service and support tasks, as shown in Figure 3.
Figure 3: Context menu showing service and support Power System actions
Both Figure 2 and Figure 3 include actions you might not be familiar with from the HMC because FSM integrates the storage, network, and higher-end virtual capabilities into one menu structure. For example, the Server to Storage Mapping View action in System Configuration provides a true mapping of storage to the virtual servers that are using it. Another example, the Submit service request action in Service and Support, is a common action across all FSM resources. It lets you submit a problem to IBM, not just for the managed Power Systems but also for the chassis, switches, IBM Storwize V7000, and other resources.
Power System Focus
Although you can manage many kinds of resources from this chassis view, there will be times you’ll want to focus entirely on your Power System resources. This is where the Manage Power System Resources tab comes in. From the chassis map, click the General Actions drop-down menu and select Manage Power System Resources. A new tab appears, and you’ll see the Explorer view shown in Figure 4.
Figure 4: Two Power System compute nodes
Once opened, this Power System view can be your home for managing the various Power System resources as well as running familiar HMC-style actions. You can customize which tabs open by default when you log on, so you can add a Manage Power System Resources tab to your default list. To add this tab, select Add to My Startup Pages. Once you log on again, the Manage Power System Resources tab will appear.
The tree at the left-hand side of Figure 4 helps you focus on essential information: the list of Power System hosts and a table of virtual servers, OSs, and power units. Not only is the list of resources dedicated to Power Systems, but the columns are optimized to show what Power System users need. For example, the list of Power System compute nodes shows available (and configurable) processors and memory reference codes, which lets you immediately see which host has room for that new AIX image you want to run.
As another example, if you select the Virtual Servers view, you’ll see a list of all virtual servers running across all compute nodes. Because you can have up to 14 Power System compute nodes per chassis and up to four chassis total, the virtual server table can be quite long. In that table, columns become even more important because you can sort and filter (and use the table search box) to find exactly what you need by entering text that matches any of the columns. The columns include partition ID, detailed status, memory, processing units, active profile, OS version, memory mode, processor mode, IPL source, state, and the IP address.
Now I’ll provide a few examples of the capabilities in FSM’s Power System resources. If you’re accustomed to using your HMC, you’ll be familiar with these capabilities.
The first example is editing a host, or managed, system. From anywhere in FSM (e.g., chassis map, list of problems, system pool dashboard, Power Systems Resources view), you can right-click any Power System compute node and view details about its properties, processors, memory, I/O, migration history, Power-On parameters, active capabilities, and more, as shown in Figure 5.
Figure 5: Using Edit Host to view details about the state of a Power System
Using FSM, you can manage a virtual server and edit many of the settings you expect, as Figure 6 shows.
Figure 6: Editing a virtual server
Of course, you’ll need to power off the virtual server to be able to edit this many parameters, but you can still edit a few parameters when the virtual server is powered on. Notice in Figure 6 that even though you’re digging deep into PowerVM capabilities, the banner at the top of the UI still keeps you posted on any critical issues that might show up for the chassis, switches, storage systems, or even a system pool managing all the virtual workloads. The details of what’s most critical are only a single click away.
Automation and thresholds that are built into FSM are also just a click away. From the Power System Resources view, you can see the Performance Summary, shown in Figure 7.
Figure 7: Performance summary of memory usage and availability for the hosts and virtual servers
This summary provides a list of processor and memory metrics that show real-time data, letting you activate thresholds for each metric and get emails or perform automated actions when those thresholds are exceeded.
Another example of the broad reach of FSM is that in the same context menu as for editing a host, you can select Server to Storage mapping, shown in Figure 8.
Figure 8: Server to storage mapping
The screen in Figure 8 shows just a snippet of what you can see: a list of all your virtual servers and where the storage is coming from. Notice that the Server to Storage display shows the virtual server, the virtual server volume, the VIOS, the Fibre Channel port, the worldwide port name, the switches, and the storage zoning information all in one table.
Once you have your Power System resources working the way you want, you can explore other areas of FSM, such as VMControl. VMControl lets you pool your Power System compute nodes together into a single pool of servers, then deploy ready-to-use virtual appliances into the system pool. The system pool is built with policies, so all you need to do is choose which virtual appliance (or image) to deploy, select the system pool (if you have more than one), and click Finish. The system pool then takes over. While you’re doing other work, the system pool will perform these tasks:
- Analyze the virtual appliance to find how much processor and memory it needs to thrive.
- Determine what networking is needed.
- Decide which compute node in the system pool has the resources available to run the virtual appliance.
- Analyze how many disks are built into the virtual appliance and use storage copy services (or V7000’s FlashCopy) to quickly allocate the proper storage.
- Create the virtual server (partition) based on the virtual appliance details. • Assign virtual LAN (VLAN) settings required by the virtual appliance.
- Boot the guest OS.
- Perform initial customization to personalize the image based on the hostname and IP address you gave the virtual appliance during the deployment.
When the virtual appliance is deployed, you can then manage the newly created workload and even group virtual servers together to manage tiered workloads. (I went into quite a bit of detail on managing workloads in VMControl in “Managing Your Data Center Through Workloads.”) The result is that you can perform these cloud-enabling activities from the same UI in which you viewed the reference code history of a specific Power System compute node.
Many of the HMC-style actions I’ve used are also available via FSM, such as viewing the hardware topology, working with system plans, and launching Advanced System Management (ASM), shown in Figure 9.
Figure 9: Advanced System Management showing vital product data for this compute node
Although ASM shows up in a new browser tab, I can still manage every detail of my Power System compute node in one browser, even if I need to dive deep into ASM. Fortunately, ASM comes enabled with single sign-on, so I don’t have to be distracted when troubleshooting a problem.
FSM Provides Comprehensive Management
I’ve provided just a few examples of how the capabilities in FSM extend way beyond basic hardware management. FSM can help you manage your Power System compute nodes in a much more complete way through the Manage Power System Resources tab and the built-in, HMC-style context menu actions, giving you a much broader view of what you need to manage through integrating automation, monitoring, storage, network, and high-end workload deployment and system pool virtualization.
Learn more about IBM Flex System Manager:
Editor's Note: This article originally appeared in the October 2012 digital issue of POWER IT Pro.