Details
-
Bug
-
Status: Resolved
-
Critical
-
Resolution: Fixed
-
2.7.0
-
None
Description
HDFS NameNode Federated clusters require a "clusterId" setting during the initial NameNode formatting phase. The HDFS stack scripts use a new configuration property to allow customization of the clusterId: {code}hadoop-env/dfs_ha_initial_cluster_id{code}.
The stack changes will read in this config property, and use this as the clusterId for the HDFS cluster.
This current JIRA tracks a minor change to the BlueprintConfigurationProcessor. The config processor should be updated to check to see if {code}hadoop-env/dfs_ha_initial_cluster_id{code} has been specified by the Blueprint deployer. If the property has not been set, then the configuration processor should auto-generate this, using the cluster name as a default.
In addition to setting the cluster ID, this change should also implement some minor configuration conveniences that will make Blueprint creation for Federated clusters simpler:
1. Add property updater support for "dfs.namenode.shared.edits.dir.NAMESERVICE" properties.
2. Add property update support for "dfs.namenode.servicerpc-address.NAMESERVICE.NAMENODE" properties