AIX > Administrator > Systems Management

Working With Power Enterprise Pools

IBM Power Enterprise Pools provide support for the use of mobile processor core and memory activations within a pool of Power Systems servers. This article focuses on the creation of a Power Enterprise Pool and the operational use of mobile activations.

Preliminary Requirements

The following steps must be completed to prior to using the HMC to create and work with a Power Enterprise Pool (pool) environment:

  • Order one or more servers with mobile activations for cores and/or memory. A server needs to have mobile activations and/or unallocated Capacity on Demand (CoD) resources to be an active participant in a pool.
  • Choose a set of servers to participate in a specific pool and send their serial numbers to IBM. IBM will generate an XML file that contains information about the servers in the pool and the total quantity of mobile activations.

After these steps have been completed, the pool configuration work can begin at the HMC. Note that the HMC menus and informational tables frequently use the term processors, which should be interpreted as processor cores.

Getting Started With the HMC

Using the master HMC, select Power Enterprise Pools, which is located in the Systems Management section on the tasks pane. This will take you to the main Power Enterprise Pools screen. Click the Create Pool button to define a new pool. You’ll be prompted for a pool name and the location of the XML definition file. Once completed, the new pool will be listed on the screen along with any other pools that may have already been set up. For each pool being managed by the HMC, the following information will be displayed:

  • Pool Name
  • Compliance Status
  • Total Mobile CoD processors
  • Available Mobile CoD processors
  • Total Mobile CoD Memory (GB)
  • Available Mobile CoD Memory (GB)

Clicking on one of the pool names presents the following selectable options in the task pane:

  • Processor Resources
  • Memory Resources
  • General Pool Properties
  • Compliance Information
  • View Pool History Log

The Processor Resources option serves as the primary dashboard (panel) for mobile processor allocations within the pool. It displays a summary of the mobile processor resources for the entire pool, showing mobile processor counts for total in pool, in use and available. Detailed processor information is also provided for each server in the pool. This includes the number of processors installed in each server along with the quantity of non-mobile activations (these could be permanent or elastic on/off CoD) and the quantity of active mobile activations for that server. The Memory Resources tab provides a similar dashboard for installed, allocated, and available memory.

Adding and Removing Mobile Resources

The Processor Resources panel shows an additional column titled Active Mobile CoD Proposed. This column lists proposed changes to the mobile processor allocations for each server in the pool. If no changes are pending, the Active Mobile CoD and Active Mobile CoD Proposed counts will be the same for each server in the pool. In the Active Mobile CoD Proposed column, an Edit button for each server in the pool leads you to an input window that lets you make changes to the Active Mobile CoD Proposed counts. Increasing the proposed count will redirect mobile processor activations to the selected server. Note that the server will need to have unallocated CoD processor resources available to make use of additional mobile activations. Decreasing the Active Mobile CoD Proposed count will remove mobile processor activations from the selected server. The hypervisor won’t forcefully remove mobile activations from a server. You’ll need to be sure that that the LPAR definitions for the server are configured so that the mobile processor allocations can be removed from that server’s shared processor pool. Similar operations apply to mobile memory allocations.

Operational Compliance

Under normal conditions, the Processor Resources panel will show a compliance status of In Compliance. It’s possible, for short periods of time, to overcommit the total quantity of mobile processor activations. This may be needed for Live Partition Mobility (LPM) events. When using LPM to move an LPAR from one server to another, you must activate processor and memory resources available on the target server that are equal to the size of the LPAR on the source server. The concept of overcommitment allows you to allocate mobile resources to the target server, potentially exceeding the total quantity of mobile resources within the pool. If the quantity of mobile processor activations exceeds that total configured for the pool, the pool is considered to be overcommitted. There’s a 24-hour grace period for overcommitted resources. After the LPM event is finished, you can remove mobile resources from the source server and no longer be in an overcommitted situation. If the grace period expires for any of the servers in the pool, the pool is marked Out of Compliance. When servers are in an Out of Compliance state, you can only assign mobile processors to the servers that have unreturned mobile processor allocations. This is how you get the pool back to an In Compliance state. Compliance for mobile memory activations follows the same process.

Adding or Removing Servers From the Pool

After the initial pool is created, you may want to add additional servers to it. The list of servers to be added must be submitted to IBM so that a new XML configuration file can be generated. If you want to remove a server from the pool, all mobile activations that were purchased with that server must be returned to it, prior to its removal from the pool. You would then submit server serial numbers to IBM so that a new XML file can be generated for the pool.

Planning for the Use of Power Enterprise Pools

Having knowledge of how Power Enterprise Pools are configured and modified can help you plan for the sizing and use of mobile activations within your enterprise. The grace period is helpful for LPM events and unplanned computing peaks. For unplanned peaks, you can immediately assign mobile processor and memory activations to the server in need. The grace period provides time to later decide where to take resources from to maintain an In Compliance state.

Charlie Cler supports customers in a solutions-architect role at Forsythe Technology Inc. He can be reached at

Like what you just read? To receive technical tips and articles directly in your inbox twice per month, sign up for the EXTRA e-newsletter here.

comments powered by Disqus



2019 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.


How to Download Fixes


Understand your options for 12X PCIe I/O drawers

clmgr: A Technical Reference

PowerHA SystemMirror 7.1 introduces a robust CLI utility

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store
IBMi News Sign Up Today! Past News Letters