Before you can begin using the management pack, you must create an adapter instance to identify the host from which the management pack will retrieve data.
- From the top navigation bar, select Administration. In the right panel, the Solutions view will be displayed.
- Select Apache Hadoop from the Solutions list on the right.
- Click the Configure icon. The Manage Solution window will appear.
Note: Click the Add icon above the Instance Namelist on the left to create multiple adapter instances.
- In the Manage Solution window, enter the following information:
- Display Name: Enter a name for this particular instance of the management pack.
- Description: Optional field, but it can be helpful to describe multiple adapter instances of the management pack.
- Basic Settings:
NameNode Host: Enter the Hostname/IP address of your Hadoop NameNode JMX endpoint
ResourceManager Host: Enter the Hostname/IP address of your Hadoop ResourceManager JMX endpoint.Note: A credential is not used for the Management Pack for Apache Hadoop because JMX is read-only.
- Advanced Settings:
- Collector: The collector you select determines which vRealize Operations Manager collector is used to manage the adapter processes. Default Collector/Group is automatically selected. Click the drop-down menu if you want to run the collection on a different node.
- NameNode Port: Enter the port associated with your NameNode JMX endpoint (default is 50070).
ResourceManager Port: Enter the port associated with your ResourceManager JMX endpoint (default is 8088).
Collection Mode: Select your desired collection mode: Standard (collects everything) or Light (collects top-level NameNode and ResourceManager resources only).Note: In Light collection mode the management pack connects to top-level NameNode and ResourceManager resources only; however, it still provides select alerting for DataNode and NodeManager resources. Use of Light mode may help address performance problems by reducing the number of metric calls.
- Timeout (seconds): Specify the timeout value for each HTTP call; default value is 5 seconds.
- Max Threads: Specify the maximum number of threads to use during collection; field requires a minimum of 1 thread.
- Dead Node Delete Age (Hours): Enter the number of hours to keep nodes that are determined to be dead/unreachable before setting them to notExisting. The default value is 0, which never sets dead nodes to notExisting.
- Support Autodiscovery: Default setting is True, which enables the adapter instance to create resources for you. This should not be set to False.
- Click Test Connection to test the configured connection information.
- Click Save Settings and Close to save your adapter instance.
What to do next
- To validate data collection, see: Validating Data Collection (Apache Hadoop)