Contents

Configuring process placement 1

Overview·· 1

Process· 1

1:N process redundancy· 1

Process placement policy and optimization· 1

Configuration restrictions and guidelines· 2

Process placement configuration task list 3

Configuring process placement policy· 3

Configuring a location affinity· 3

Configuring a location type affinity· 3

Configuring a process affinity· 4

Configuring a self affinity· 4

Optimizing process placement 5

Displaying process placement 5

 

Configuring process placement

Overview

Process placement enables placing processes to specific CPUs (also called nodes) on the main processing units (MPUs) in your system for optimal distribution of CPU and memory resources.

Process

A process contains a set of codes and provides specific functionality. For example, an AAA process provides AAA functions.

Each process runs in a protected memory space to prevent problems with one process from impacting the entire system.

1:N process redundancy

The system backs up each active process running on one node to all the other nodes. When an active process fails, one of its standby processes promptly takes over without impacting any other service.

1:N process redundancy provides the following benefits:

·     Improves service availability.

·     Enables the system to quickly regain reliability after device status changes in such conditions as insertion and removal of cards, IRF split, and removal of an IRF member.

Process placement policy and optimization

Process placement policies

(In standalone mode.) The execution of process placement policies varies by the location of active processes.

·     An active process running only on the active MPU does not support placement optimization. If you configure a process placement policy for the process, the system displays a configuration failure prompt. When such an active process fails, the system automatically restarts the process. The standby processes are used for active/standby switchover and ISSU.

·     Some active processes can run on either the active or standby MPU. When such an active process fails, the system uses a placement policy to select a new active process among standby processes.

(In IRF mode.) The execution of process placement policies varies by the location of active processes.

·     An active process running only on the global active MPU does not support placement optimization. If you configure a process placement policy for the process, the system displays a configuration failure prompt. When such an active process fails, the system automatically restarts the process. The standby processes are used for active/standby switchover and ISSU.

·     Some active processes can run on either the global active or standby MPU. When such an active process fails, the system uses a placement policy to select a new active process among standby processes.

The system provides a default process placement policy that takes effect for all processes. You can modify the default placement policy in the view you enter by using the placement program default command. You can also configure a placement policy for a specific process in the view you enter by using the placement program program-name [ instance instance-name ] command. A placement policy for a process takes precedence over the default process placement policy.

By default, the default process placement policy defines the following rules:

·     (In standalone mode.) The active process runs on the main CPU of the active MPU, and the standby processes run on the main CPU of the standby MPU.

·     (In IRF mode.) The active process runs on the main CPU of the global active MPU, and the standby processes run on the main CPU of the global standby MPU.

·     A process runs at the location where it ran the last time and does not move to any other location during startup or operation.

·     The addition of a new node does not impact current active processes. A new active process selects one node with sufficient CPU and memory resources. (You can use the display cpu-usage and display memory commands to view CPU and memory usage information.)

Optimizing process placement

You can configure the following settings for a process placement policy to optimize process placement:

·     affinity location-set—Location affinity, the preference for the process to run on a specific node.

·     affinity location-type—Location type affinity, the preference for the process to run on a particular type of node. For more information about node types, see "Configuring a location type affinity."

·     affinity program—Process affinity, the preference for the process to run on the same node as a particular process.

·     affinity self—Self affinity, the preference for one instance of the process to run on the same node as any other instance of the process.

Affinities include positive affinities (attract) and negative affinities (repulse), all represented by integers in the range of 1 to 100000.

·     The higher the attract value, the stronger the preference.

·     The higher the repulse value, the weaker the preference.

After you apply new placement policies, the system makes placement decisions based on the new policies, node resources, and topology status. If the new location for an active process is different from the current node, the system changes the state of the process to standby, and uses the standby process on the preferred location as the new active process.

Configuration restrictions and guidelines

When you configure process placement, follow these restrictions and guidelines:

·     Configuring process placement on a device with only one MPU does not change the location of processes. All processes run on the main CPU of the MPU.

·     Configuring process placement on a device with multiple MPUs places specific active processes to specific CPUs. In case of multiple CPUs, the system performs 1:N process redundancy. The number of standby processes and their CPU locations vary by function module. The system by default automatically determines the location for each active process, and process placement optimization is not required. If optimization is required, contact H3C Support to avoid service interruption.

·     Do not perform process placement for a process frequently.

·     Process placement applies only to MPUs.

·     To view the current location of an active process and its predicted new location after optimization, use the display placement reoptimize command.

Process placement configuration task list

Tasks at a glance
Configuring process placement policy: ·     (Optional.) Configuring a location affinity ·     (Optional.) Configuring a location type affinity ·     (Optional.) Configuring a process affinity ·     (Optional.) Configuring a self affinity
(Required.) Optimizing process placement

 

Configuring process placement policy

Configuring a location affinity

Step	Command	Remarks
1.     Enter system view.	system-view	N/A
2.     Enter placement process view.	·     Enter default placement process view: placement program default ·     Enter placement process view: placement program { program-name [ instance instance-name ]	Settings in default placement process view take effect for all processes. Settings in placement process view take effect only for the specified process.
3.     Set the location affinity.	·     In standalone mode: affinity location-set slot slot-number&<1-5> { attract strength | default | none | repulse strength } ·     In IRF mode: affinity location-set chassis chassis-number slot slot-number&<1-5> { attract strength | default | none | repulse strength }	By default, no location affinity is set.

 

Configuring a location type affinity

The following location types are available:

·     current—Current location of the active process, which can be displayed with the display placement program command.

·     paired—Locations of standby processes.

·     (In standalone mode.) primary—Active MPU.

·     (In IRF mode.) primary—Global active MPU.

To configure a location type affinity:

 

Step	Command	Remarks
1.     Enter system view.	system-view	N/A
2.     Enter placement process view.	·     Enter default placement process view: placement program default ·     Enter placement process view: placement program { program-name [ instance instance-name ]	Settings in default placement process view take effect for all processes. Settings in placement process view take effect only for the specified process.
3.     Set the location type affinity.	affinity location-type { current | paired | primary } { attract strength | default | none | repulse strength }	By default, no location type affinity is set.

 

Configuring a process affinity

Step	Command	Remarks
1.     Enter system view.	system-view	N/A
2.     Enter placement process view.	·     Enter default placement process view: placement program default ·     Enter placement process view: placement program { program-name [ instance instance-name ]	Settings in default placement process view take effect for all processes. Settings in placement process view take effect only for the specified process.
3.     Configure the affinity for the process to run on the same location as another process.	affinity program program-name { attract strength | default | none | repulse strength }	By default, no process affinity is set.

 

Configuring a self affinity

Perform this task to configure the preference for one instance of a process to run on the same node as any other instance of the process. The self affinity setting does not take effect for a process that has only one instance.

To configure a self affinity:

 

Step	Command	Remarks
1.     Enter system view.	system-view	N/A
2.     Enter placement process view.	·     Enter default placement process view: placement program default ·     Enter placement process view: placement program { program-name [ instance instance-name ]	Settings in default placement process view take effect for all processes. Settings in placement process view take effect only for the specified process.
3.     Configure a self affinity.	affinity self { attract strength | default | none | repulse strength }	By default, no self affinity is set.

 

Optimizing process placement

To avoid neighbor flapping of related protocols, make sure HA features such as NSR or GR are configured for the processes and are stable before optimizing process placement.

To optimize process placement:

 

Step	Command	Remarks
1.     Enter system view.	system-view	N/A
2.     Optimize process placement.	placement reoptimize	To keep the system stable, do not perform any tasks that require process restart when you execute this command.

 

Displaying process placement

Execute display commands in any view.

 

Task	Command
Display process placement policy information.	display placement policy program { program-name | all | default }
Display the location of a process.	display placement program { program-name | all }
(In standalone mode.) Display the running processes on a specific location.	display placement location { slot slot-number | all }
(In IRF mode.) Display the running processes on a specific location.	display placement location { chassis chassis-number slot slot-number | all }
Display the predicted location of a process after process placement optimization.	display placement reoptimize program { program-name [ instance instance-name ] | all }
Display service group information.	display ha service-group { program-name [ instance instance-name ] | all }