This is the documentation for Cloudera Manager 4.8.3.
Documentation for other versions is available at Cloudera Documentation.

Rolling Restart

Rolling restart allows you to conditionally restart the role instances of your HDFS, MapReduce, HBase, ZooKeeper, and Flume services. Note that if the service is not running, Rolling Restart is not available for that service.

  Note: This feature is available only with a Cloudera Enterprise license.
For other products, the following applies:
  • Cloudera Standard - the feature is not available.
  • Cloudera Enterprise Trial Edition - the feature will not be available after your trial license expires.
To obtain a license for Cloudera Enterprise, please fill in this form or call 866-843-7207. After you install your Cloudera Enterprise license, the feature will be available.

You can do a rolling restart of each of these services individually.

If you have HDFS High Availability enabled, you can also perform a cluster-level rolling restart. At the cluster level, the rolling restart of slave nodes is performed on a host-by-host basis, rather than per service, to avoid all roles for a service potentially being down at the same time. You cannot perform a cluster-level rolling restart unless you have High Availability enabled. However, you can perform a rolling restart on individual services, including HDFS. Note that JobTracker High Availability is not required for a cluster-level rolling restart.

If you have HDFS High Availability, then in order to avoid having your NameNode (and thus the cluster) go down during the restart, Cloudera Manager will force a failover to the Standby NameNode while the critical roles are being restarted. If you have JobTracker High Availability enabled, Cloudera Manager will force a failover to the Standby JobTracker while the critical roles are being restarted.

Restarting an Individual Service

You can initiate a rolling restart from either the Service page for one of the eligible services, or from the service's Instances page, where you can select individual roles to be restarted.

  1. From the Services page (or the Services tab) select the service you want to restart.
  2. From the service's Actions menu, select Rolling Restart...

    OR:

    1. Go to the Instances tab.
    2. Select the roles you want to restart.
    3. Select Rolling Restart from the Actions for Selected menu.
  3. In the pop-up dialog box, select the options you want:
    • You can choose to restart only roles who's configurations are stale
    • You can choose to restart only roles that are running outdated software versions.
    • Select which role types you want to restart.
  4. If you have a significant number of slave roles (DataNodes, TaskTrackers, RegionServers) you can have those restarted in batches. You can configure:
    • How many roles should be included in a batch (the default is one, so individual roles will be started one at a time).
    • How long should Cloudera Manager wait before starting the next batch.
    • The number of batch failures that will cause the entire rolling restart to fail (this is an advanced feature).
  5. Click Confirm to start the Rolling Restart.

Restarting HDFS

If you do not have HDFS High Availability configured, a warning appears reminding you that the service will become unavailable during the restart while the NameNode is restarted. Services that depend on that HDFS service will also be disrupted.

It is recommended that you restart the DataNodes one at a time — one host per batch, which is the default.

Restarting HBase

Administration operations such as any of the following should not be performed during the rolling restart, to avoid leaving the cluster in an inconsistent state:
  1. Splits
  2. Create/disable/enable/drop table
  3. Metadata changes
  4. Create, clone, or restore a snapshot. Snapshots rely on the Region Servers being up; otherwise the snapshot will fail.

Restarting MapReduce

If you restart the JobTracker, and Hall current jobs will fail.

Restarting ZooKeeper or Flume

For both ZooKeeper and Flume, the option to restart roles in batches is not available. They are always restarted one by one.

Restarting a Cluster

You can perform a cluster-level rolling restart on demand (from the Cloudera Manager admin console). A cluster-level rolling restart is also performed as the last step in a rolling upgrade when the cluster is configured with HDFS High Availability enabled.

  Note:

Rolling Restart for a cluster is available ONLY if you have HDFS High Availability enabled. In order to avoid having your NameNode go down during the restart, Cloudera Manager will force a failover to the Standby NameNode while the critical roles are being restarted. JobTracker High Availability is not required, although Cloudera Manager will also force a failover if the feature is enabled.

  1. If you have not already done so, enable High Availability. See Configuring HDFS High Availability for instructions. You do not need to enable Automatic Failover for rolling restart to work, though you can enable it if you wish. Automatic Failover does not affect the rolling restart operation.
  2. From the Actions menu for the cluster you want to restart (found on the All Services page) select Rolling Restart....
  3. In the pop-up dialog box, select the services you want to restart. Please review the caveats in the preceding sections for the services you elect to have restarted. Note that the services that do not support rolling restart will simply be restarted, and will be unavailable during their restart.
  4. If you select an HDFS, HBase, or MapReduce service, you can also configure the following:
    • How many slave hosts should be included in a batch (the default is one, so slave roles will be restart host by host).
    • How long should Cloudera Manager wait before starting the next batch.
    • The number of batch failures that will cause the entire rolling restart to fail (this is an advanced feature).
  5. Click Confirm to start the Rolling Restart. While the restart is in progress, the Command Details page shows the steps for stopping and restarting the services.