Configuring Cloudera Manager Agents

Minimum Required Role: Full Administrator

Cloudera Manager Agents can be configured globally using properties you set in the Cloudera Manager Admin Console and by setting properties in Agent configuration files.

Continue reading:

Configuring Agent Heartbeat and Health Status Options

You can configure the Cloudera Manager Agent heartbeat interval and timeouts to trigger changes in Agent health as follows:

  1. Select Administration > Settings.
  2. Under the Performance category, set the following option:
    Property Description
    Send Agent Heartbeat Every The interval in seconds between each heartbeat that is sent from Cloudera Manager Agents to the Cloudera Manager Server.

    Default: 15 sec.
  3. Under the Monitoring category, set the following options:
    Property Description
    Set health status to Concerning if the Agent heartbeats fail The number of missed consecutive heartbeats after which a Concerning health status is assigned to that Agent.

    Default: 5.
    Set health status to Bad if the Agent heartbeats fail The number of missed consecutive heartbeats after which a Bad health status is assigned to that Agent.

    Default: 10.
  4. Click Save Changes.

Configuring the Host Parcel Directory

To configure the location of distributed parcels:
  1. Click Hosts in the top navigation bar.
  2. Click the Configuration tab.
  3. Select Category > Parcels.
  4. Configure the value of the Parcel Directory property. The setting of the parcel_dir property in the Cloudera Manager Agent configuration file overrides this setting.
  5. Enter a Reason for change, and then click Save Changes to commit the changes.
  6. Restart the Cloudera Manager Agent on all hosts.

Agent Configuration File

The Cloudera Manager Agent supports different types of configuration options in the /etc/cloudera-scm-agent/config.ini file. You must update the configuration on each host. After changing a property, restart the Agent: 
sudo service cloudera-scm-agent restart
Section Property Description
[General] server_host, server_port, listening_port, listening_hostname, listening_ip Hostname and ports of the Cloudera Manager Server and Agent and IP address of the Agent. Also see Configuring Cloudera Manager Server Ports and Ports Used by Cloudera Manager and Cloudera Navigator.

The Cloudera Manager Agent configures its hostname automatically. You can also manually specify the hostname the Cloudera Manager Agent uses by updating the listening_hostname property. To manually specify the IP address the Cloudera Manager Agent uses, update the listening_ip property in the same file.

To have a CNAME used throughout instead of the regular hostname, an Agent can be configured to use listening_hostname=CNAME. In this case, the CNAME should resolve to the same IP address as the IP address of the hostname on that machine. Users doing this will find that the host inspector will report problems, but the CNAME will be used in all configurations where that's appropriate. This practice is particularly useful for users who would like clients to use namenode.mycluster.company.com instead of machine1234.mycluster.company.com. In this case, namenode.mycluster would be a CNAME for machine1234.mycluster, and the generated client configurations (and internal configurations as well) would use the CNAME.
lib_dir Directory to store Cloudera Manager Agent state that persists across instances of the agent process and system reboots. The Agent UUID is stored here.

Default: /var/lib/cloudera-scm-agent.
local_filesystem_whitelist The list of local filesystems that should always be monitored.

Default: ext2,ext3,ext4.
log_file The path to the Agent log file. If the Agent is being started using the init.d script, /var/log/cloudera-scm-agent/cloudera-scm-agent.out will also have a small amount of output (from before logging is initialized).

Default: /var/log/cloudera-scm-agent/cloudera-scm-agent.log.
max_collection_wait_seconds Maximum time to wait for all metric collectors to finish collecting data.

Default: 10 sec.
metrics_url_timeout_seconds Maximum time to wait when connecting to a local role's web server to fetch metrics.

Default: 30 sec.
parcel_dir Directory to store unpacked parcels.

Default: /opt/cloudera/parcels.

If you want to change this, Cloudera recommends following the procedure documented in Changing the Parcel Directory to change this for all hosts, rather than setting it in each host config.ini file.

This property overrides the setting in Cloudera Manager. To use the recommended procedure, you must make sure that this property is commented out in each host config.ini file.
supervisord_port The supervisord port. A change takes effect the next time supervisord is restarted (not when the Agent is restarted).

Default: 19001.
task_metrics_timeout_seconds Maximum time to wait when connecting to a local TaskTracker to fetch task attempt data.

Default: 5 sec.
[Security] use_tls,verify_cert_file, client_key_file, client_keypw_file, client_cert_file Security-related configuration. See
[Cloudera] mgmt_home Directory to store Cloudera Management Service files.

Default: /usr/share/cmf.
[JDBC] cloudera_mysql_connector_jar, cloudera_oracle_connector_jar, cloudera_postgresql_jdbc_jar Location of JDBC drivers. See Step 4: Install and Configure Databases.
Default:
  • MySQL - /usr/share/java/mysql-connector-java.jar
  • Oracle - /usr/share/java/oracle-connector-java.jar
  • PostgreSQL - /usr/share/cmf/lib/postgresql-version-build.jdbc4.jar