Deploying Client Configuration Files

To allow Hadoop client users to work with the HDFS, MapReduce, YARN and HBase services you created, Cloudera Manager creates zip files that contain the relevant configuration files with the settings for your services. Each zip file contains the set of configuration files needed by the appropriate service: for example, the MapReduce client configuration zip file contains copies of core-site.xml,, hdfs-site.xml, and mapred-site.xml.

These client configuration files are generated automatically by Cloudera Manager based on the services and roles you have installed.

Cloudera Manager deploys these configurations automatically when you install your cluster, when you add a service on a host, or when you add a Gateway role on a host. Specifically, for each host that has a service role instance installed, and for each host that is configured as a Gateway role for that service, the Deploy function downloads the configuration zip file, unzips it into the appropriate configuration directory, and uses the Linux "alternatives" mechanism to set a given, configurable priority level.


A Gateway is a role whose sole purpose is to designate a host that should receive a client configuration for a specific service, when the host does not otherwise have any roles running on it. Gateways are configured by going to the Instances tab for the service and using the Add command to add Gateway roles as needed. You can configure Gateway roles for HDFS, MapReduce, and HBase services (and for YARN in CDH4). See Adding Role Instances for more information on adding Gateway roles.

Note that if you are installing on a system that happens to to have pre-existing alternatives, then it is possible another alternative may have higher priority and will continue to be used. The alternatives priority of the Cloudera Manager client configuration is configurable under the Client section of the Configuration tab for the appropriate service.

You can also distribute these client configuration files manually to the users of a service.

The main circumstance that may require a redeployment of the client configuration files is when you have modified the configuration of your cluster. In this case you will typically see a message telling you to redeploy your client configurations. The affected service(s) will also typically be shown as "Running with outdated Configuration."

Viewing and Downloading the Client Configuration Files

You can view the client configuration files using the Client Configuration URLs button from the main Services tab:

  1. Click the Services tab in the Cloudera Manager Admin Console.
  2. Click the Client Configuration URLs button. This opens a popup window with links to the configuration zip files that have been created for the services you have installed: HDFS, MapReduce, YARN, and HBase.
  3. Click a link to initiate a download of the configuration zip file to your local system.

To download an individual client configuration zip file:

You can download client configuration files for HDFS, MapReduce, YARN, and HBase.

  1. Click the Services tab in the Cloudera Manager Admin Console and select the service instance whose configuration you want to download.
  2. From the Actions menu at the top right of the service page, select Download Client Configuration This initiates a download to your local system of the configuration files for the selected service.

The client configuration files can be downloaded without authentication using tools like wget and curl.

Redeploying the Client Configuration Files Manually

Although Cloudera Manager will deploy client configuration files automatically in many cases, if you have modified the configurations for a service, you may need to redeploy those configuration files.

If your client configurations were deployed automatically, this command will attempt to redeploy them as appropriate.


If you are deploying client configurations on a node that has multiple services installed, some of the same configuration files, though with different configurations, will be installed in the conf directories for each service. Cloudera Manager uses the priority parameter in the alternatives --install command to ensure that the correct configuration directory is made active based on the combination of services on that node. The priority order (as of Cloudera Manager 4.1.2) is MapReduce > YARN > HDFS. The priority can be configured under the Client/Advanced section of the Configuration tab for the appropriate service.

To deploy all the client configuration files to all nodes on your cluster:

  1. Click the Services tab in the Cloudera Manager Admin Console.
  2. From the cluster-level Actions menu at the top right of the page, select Deploy Client Configuration...
  3. If you are sure you want to proceed, click Deploy client configuration.

To deploy client configuration files for a specific service:

  1. From the Services tab, click the service for which you want to deploy client configurations.
  2. From the Actions menu at the top right of the service page, select Deploy client Configuration...
  3. If you are sure you want to proceed, click Deploy client configuration.

How Client Configurations are Deployed

Client configuration files are deployed on any host that is a client for a service — i.e. that has a role for the service on that host. This includes roles such as DataNodes, TaskTrackers, RegionServers and so on as well as Gateway roles for the service.

If roles for multiple services are running on the same host (e.g. a DataNode role and a TaskTracker role on the same host) then the client configurations for both roles are deployed on that host, with the alternatives priority determining which configuration takes precedence.

For example, if we have six hosts running roles as follows: Host H1: HDFS-NameNode Host H2: MR-JobTracker Host H3: HBase-Master Host H4: MR-TaskTracker, HDFS-DataNode, HBase-RegionServer Host H5: MR-Gateway Host H6: HBase-Gateway

Client configuration files will be deployed on these hosts as follows: Host H1: hdfs-clientconfig (only) Host H2: mapreduce-clientconfig Host H3: hbase-clientconfig Host H4: hdfs-clientconfig, mapreduce-clientconfig, hbase-clientconfig Host H5: mapreduce-clientconfig Host H6: hbase-clientconfig

If the HDFS NameNode and MR JobTracker were on the same host, then that host would have both hdfs-clientconfig and mapreduce-clientconfig installed.