Configuring AX Engine Nodes
Use one or more instances of AX Engine Node to run analytics across multiple servers and increase processing power.
How it works
Increasing processing power
By default, analytics are processed directly on AX Server. Although this works in most cases, you may require more processing power when:
- you want to run anlaytics requiring high processing power
- you want to run a number of analytics in a short span of time
To increase processing power, you can configure one or more instances of AX Engine Node and run your analytics on these separate servers.
Note
When AX Engine Node is enabled, analytic jobs do not run on AX Server. All jobs run on an instance of AX Engine Node.
Balancing multiple nodes
When a node has space in its queue, it notifies AX Server that it can process jobs and the server then assigns jobs to that node. As a result, running multiple AX Engine Nodes that are allowed to handle several jobs simultaneously can cause one node to handle multiple jobs while others sit idle.
To ensure jobs are balanced across nodes, configure the maximum number of jobs each node can process as 1 and schedule analytic jobs accordingly.
Setting up AX Engine Nodes
Prerequisites
Before you add and configure an AX Engine Node in the AX Server Configuration web application, you must:
- configure your system so that AX Server data files are stored in a shared folder which is accessible using a UNC path (e.g. \\server_name\shared_folder). For more information, see Configuring a shared data folder
- physically install and configure each instance of AX Engine Node on a server
Add an AX Engine Node
- Sign in to the AX Server Configuration web application and locate the Engine Nodes panel.
- In the Add Engine Node panel, enter the following information:
- New Engine Node IP address or computer name: the hostname or IP address of the server where the AX Engine Node is installed
- Jobs (max): the maximum number of analytic jobs that the engine node can process at one time
Note
Jobs are queued until the number of concurrently running jobs falls below the maximum.
- Click Add Engine Node.
- Click Update Engine Node Settings.
Result: The AX Engine Node is used to process analytics. If it is the first AX Engine Node, processing is transferred from AX Server to the new node. If you added multiple AX Engine Nodes, scheduled jobs are distributed among them.
Configuring AX Engine Nodes
From the Engine Nodes panel of the AX Server Configuration web application, you can edit the configuration settings of any AX Engine Node you have added.
When you edit a value, click Update to save the change.
Individual AX Engine Node settings
- Enabled/disable an engine node
- Change the Hostname or IP address
- Change the max jobs
- Remove an engine node
Global AX Engine Node settings
Note
These settings apply to all instances of AX Engine Node and all analytics that run on the server(s).
You can copy data files to each AX Engine Node instance before analytics are processed using the Copy analytic data to engine node setting:
- Yes: data files from the analytic folder of AX Server are copied to the analytic job directory of the AX Engine Node before processing starts
- No: data files remain in the analytic folder of AX Engine Node and are not copied to the job directory of the AX Engine Node before processing starts
| Copy data files | Do not copy data files |
|---|---|
| Improves performance when data files are repeatedly accessed across the network | Improves performance when a small number of analytics run simultaneously |
| Improves performance when running multiple analytics simultaneously or if a single analytic runs several commands on large data files |
Improves performance when the time to copy data files outweighs the time you can save by by writing your analytics so that, whenever possible, analytic commands are run locally on the AX Engine Node. For example:
|
