Power Integrated Facility for Linux
Power Integrated Facility for Linux (Power IFL) is an optional lower cost per processor core activation feature for Linux only workloads on IBM Power Systems servers. Processor cores that are activated for general-purpose workloads can run any supported operating system. If you choose to activate Power IFL processor cores, the systems need to be in compliance with the license terms.
Additional information on new HMC features can be found in IBM Power Systems Hardware Management Console Version 8 Release 8.1.0 Enhancements, SG24-8232 available at:
http://www.redbooks.ibm.com/redpieces/abstracts/sg248232.html?Open
What is new in Power IFL
Since the previous update, changes have been introduced in Hardware Management Console (HMC) Version 8 Release 8.1 to assist with managing the compliance of Power IFL processors.
The changes are:
-
Enabled Power IFL processors can be viewed from the HMC Graphical User Interface
-
Updated Command Line Interface (CLI) commands
-
Updated compliance monitoring assistance feature
Command line and graphical interface updates
Power IFL was introduced in HMC 7.9.0 and only had command line tools for monitoring the activated processor allocation and activation.
In HMC V8.8.1.0 the Capacity on Demand (CoD) Processor Capacity Settings and managed system properties in the Graphical User Interface (GUI) have been updated to show the activations and enable the monitoring of Power IFL processor allocation and activation.
The CLI commands have also been updated to show IFL activations and available IFL processor cores.
Capacity on Demand CLI and graphical interface changes
The lscode command displays the permanent Linux only and all Operating System processors. In Example 0-1 the syntax of the lscod command and the output is displayed.
The syntax of lscode command syntax and output is as follows
lscod -t cap -c cuod -r proc -m <managed system>
perm_procs=10,perm_procs_linux=3,perm_procs_all_os=7
The perm_procs_linux=3 parameter indicates that three processor cores are licensed for Linux only workloads.
Note: An additional -F flag is required if those values are not displayed in the output.
If perm_procs_linux is 0, it will not be displayed in the default output. It will only be displayed when -F is specified.
If perm_procs_all_os = perm_procs, perm_procs_all_os will not be displayed in the default output. It will only be displayed when -F is specified.
If the managed system does not support Power IFL compliance monitoring, these attributes are not shown.
|
Support for Power IFL has been added in HMC V8.8.1.0 to easily show information about CoD Capacity Processor Settings as shown in Figure 0-1.

Figure 0-1 CoD Processor Capacity Settings showing activated IFL processors
The lshwres command has been updated to show the number of processor units that are configurable for either Linux only or all Operating System workloads as shown in Example 0-2.
Example 0-2Syntax of lshwres command and output.
lshwres -m <managed system> -r proc --level sys configurable_sys_proc_units=10.0,curr_avail_sys_proc_units=1.0,pend_avail_sys_proc_units=0.0
installed_sys_proc_units=16.0,deconfig_sys_proc_units=0,min_proc_units_per_virtual_proc=0.05,max_virtual_procs_per_lpar=256,
max_procs_per_lpar=256,max_curr_virtual_procs_per_aixlinux_lpar=64,max_curr_virtual_procs_per_vios_lpar=64,
max_curr_virtual_procs_per_os400_lpar=64,max_curr_procs_per_aixlinux_lpar=64,max_curr_procs_per_vios_lpar=64
,max_curr_procs_per_os400_lpar=64,max_shared_proc_pools=64,configurable_sys_proc_units_all_os=7.0,configurable_sys_proc_units_linux=3.0
The configurable_sys_proc_units_linux=3.0 parameter indicates that 3.0 processor cores are configured for Linux only workloads.
Note: An additional -F flag is required if the expected values are not displayed in the output.
If configurable_sys_proc_units_linux is 0, it will not be displayed in the default output. It will only be displayed when -F is specified.
If configurable_sys_proc_units_all_os = configurable_sys_proc_units, configurable_sys_proc_units_all_os will not be displayed in the default output. It will only be displayed when -F is specified.
If the managed system does not support Power IFL compliance monitoring, these attributes are invalid.
|
The updated managed system properties tab now displays the number of Linux only and any Operating System processors that have been licensed in the system as shown in Figure 0-2.

Figure 0-2 GUI window output to display the Linux-only and all OS processors.
Compliance monitoring assistance
For certain models of IBM Power Systems servers, the HMC displays a message if the managed system is not in compliance with the Power IFL license terms.
Compliance monitoring assistance is available on following models with firmware version 7.8.1, or later.
-
9119-FHB
-
9117-MMD
-
9179-MHD
System firmware on supported models periodically computes the actual processor core consumption.
Note: If your system is determined to be out of compliance with the processor core license terms, the HMC displays a message every hour. You must be logged in to the HMC GUI to see these messages. Otherwise, they are discarded.
|
On HMC V8.8.1.0, or later, you can see the license configuration for a managed server with Power IFL activations in the HMC server properties Processors tab.
Two categories are listed in the Configurable section such as:
-
Processors that are listed as Linux-only represent the number of Power IFL processor cores.
-
Processors that are listed as Any can be used for any (general purpose) workload.
This same information is available in the CoD Processor Capacity Settings window.
Compliance conditions
The system records an entry in the CoD history log when an out of compliance condition is first detected.
When the number of out of compliance processor units changes, an A7004735 system reference code (SRC) is logged.
If a system is out of compliance for 24 continuous hours, an A7004736 SRC is logged as a serviceable event.
If you determine that your system is out of compliance, you must correct the problem. Reduce the processor usage of one or more of the running AIX, IBM i, or Virtual I/O Server partitions on the managed system. Reduce processor usage through dynamic partitioning, or shut down or suspend a partition.
