This is the documentation for Cloudera Manager 5.0.x. Documentation for other versions is available at Cloudera Documentation.

Viewing Service Instance Details

  1. Do one of the following:
    • On the Home page, choose a cluster and click a service name.
    • Select Clusters > ClusterName > ServiceName.
  2. Click the Instances tab on the Services navigation bar. This shows all instances of all role types configured for the selected service.

You can also go directly to the Instances page to view instances of a specific role type by clicking one of the links under the Role Counts column. This will show only instances of the role type you selected.

The Instances page displays the results of the configuration validation checks it performs for all the role instances for this service.

  Note: The information on this page is always the Current information for the selected service and roles. This page does not support a historical view: thus, the Time Range Selector is not available.

The information on this page shows:

  • The name of the role instance. Click the name to view the role status for that role.
  • The host on which it is running. Click the host name to view the host status details for the host.
  • The rack assignment.
  • The status — a single value summarizing the state and health of the role instance.
  • The state — whether the role instance is starting, stopping, running, stopped, and so on.
  • The health — whether the role instance satisfies certain management criteria for health.
  • Whether the role is currently in maintenance mode. If the role has been set into maintenance mode explicitly, you will see the following icon (). If it is in effective maintenance mode due to the service or its host having been set into maintenance mode, the icon will be this ().
  • Whether the role is currently decommissioned.

You can sort or filter the Instances list by criteria in any of the displayed columns.

To sort the Instances list:

  1. Click the column header by which you want to sort. A small arrow indicates whether the sort is in ascending or descending order.
  2. Click the column header again to reverse the sort order.

To filter the Instances list:

  • To filter by Role, Status, Health, Decommissioned, or Maintenance Mode, select the value from the drop-down search field at the top of the column.
  • To filter by Host or Rack, type the filter value in the search field.

From the Actions for Selected menu you can stop, start, restart, or delete a role, put a role into or remove it from maintenance mode, and (for HDFS or HBase roles only) decommission or recommission a role.

To take an action on one or more roles:

  1. Check the checkbox next to the role instance(s) you want to act upon (or check the box to the top of the list to select all role instances).
  2. From the Actions for Selected menu, select the appropriate action. See Managing Services for details on these actions. (The Decommission action only applies to HDFS DataNodes, MapReduce TaskTrackers, YARN NodeManagers, and HBase RegionServers.)

To add a role instance:

  • Click the Add button. This takes you to the Add Role Instances page. See Role Instances for further information.

Role Instance Reference

The following tables contain reference information on the status, role state, and health columns for role instances.

Table 1. Status
Indicator Status Description
Started with outdated configuration For a service, this indicates the service is running, but at least one of its roles is running with a configuration that does not match the current configuration settings in Cloudera Manager. For a role, this indicates a configuration change has been made that requires a restart, and that restart has not yet occurred. Click the indicator to display the Stale Configurations page.
Starting or stopping The entity is starting up but is not yet running or the service or role is stopping but has not stopped yet.
Stopped The entity is stopped, as expected.
Down The entity is not running, but it is expected to be running.
History not available The application is in historical mode, and the entity does not have historical monitoring support. This is the case for services other than HDFS, MapReduce and HBase such as ZooKeeper, Oozie, or Hue .
Status not available The entity is not started or stopped in the same way as a regular service or role. Examples are the HDFS Balancer (which runs from the HDFS Rebalance action) or Gateway roles. The Start and Stop commands are not applicable to these instances.
None The entity does not have a status. For example, it is not something that can be running and it cannot have health.
Good health The entity is running with good health. For a specific health check, the returned result is normal or within the acceptable range. For a role or service, this means all health checks for that role or service are Good.
   Concerning health The entity is running with concerning health. For a specific health check, the returned result indicates a potential problem. Typically this means the test result has gone above (or below) a configured Warning threshold. For a role or service, this means that at least one health check is Concerning.
Bad health The entity is running with bad health. For a specific health check, the check failed, or the returned result indicates a serious problem. Typically this means the test result has gone above (or below) a configured Critical threshold. For a role or service, this means that at least one health check is Bad.
Disabled health The entity is running, but all of its health checks are disabled.
Unknown health The entity is running, but there is not enough information to determine its health.
Unknown Status of a service or role or service instance is unknown. This can occur for a number of reasons, such as the Service Monitor is not running, or connectivity to the agent doing the health monitoring has been lost.
Table 2. Role State

Icon

State

Description

Started

The role instance is running normally.

Stopped

The role instance is stopped.

Started

Indicates a configuration change has been made that requires a restart, and that restart has not yet occurred.

Starting and Stopping

The role instance is starting up but is not yet running or the service or role instance is stopping: a stop command has been issued, but the instance has not finished shutting down.

History Unavailable

Cloudera Manager does not support historical information for this role. This is the case for services such as ZooKeeper, Oozie, or Hue — services other than HDFS, MapReduce and HBase.

Not applicable

The role is not started or stopped in the same way as a regular service or role. Examples are the HDFS Balancer (which runs from the HDFS Rebalance action) or Gateway roles. The Start and Stop commands are not applicable to these instances.

Table 3. Health

Icon

 Health

Description

Good

For a specific health check, the returned result is normal or within the acceptable range.

 

Concerning

For a specific health check, the returned result indicates a potential problem. Typically this means the test result has gone above (or below) a configured Warning threshold.

Bad

For a specific health check, the check failed, or the returned result indicates a serious problem. Typically this means the test result has gone above (or below) a configured Critical threshold.

Unknown

For a role or service instance, its health is Unknown. This can occur for a number of reasons, such as the Service Monitor is not running, or connectivity to the agent doing the health monitoring has been lost.

Checks disabled

Health Checks have been disabled in the configuration for this service or role instance.

History unavailable

Cloudera Manager does not support the collection of historical information for this role or service instance. This is the case for services such as ZooKeeper, Oozie, or Hue — services other than HDFS, MapReduce and HBase.

Page generated September 3, 2015.