To ensure that a multi-node cluster remains available, configure and
enable High Availability. Configuring High Availability includes defining locations and
names of hosts in a cluster that are available to act as journal nodes and a standby
name node in the event that the primary namenode fails. To configure High Availability,
add the following properties to clusterproperties.txt
and set
values as follows:
Configuring High Availability in a Windows-based Cluster
Property | Description | Example Value | Mandatory/Optional |
---|---|---|---|
HA | Whether to deploy a highly available NameNode or not. | yes or no | Optional |
NN_HA_JOURNALNODE_HOSTS | A comma-separated list of FQDN for those cluster nodes that will run the JournalNode processes. | journalnode1.acme.com, journalnode2.acme.com,
journalnode3.acme.com | Optional |
NN_HA_CLUSTER_NAME | This name is used for both configuration and authority component of absolute HDFS paths in the cluster. |
hdp2-ha | Optional |
NN_HA_JOURNALNODE_EDITS_DIR | This is the absolute path on the JournalNode machines where the edits and other local state used by the JournalNodes (JNs) are stored. You can only use a single path for this configuration. |
d:\hadoop\journal
| Optional |
NN_HA_STANDBY_NAMENODE_HOST
| The host for the standby NameNode. | STANDBY_NAMENODE.acme.com | Optional |
RM_HA_CLUSTER_NAME | A logical name for the Resource Manager cluster. | HA Resource Manager | Optional |
RM_HA_STANDBY_RESOURCEMANAGER_HOST | The FQDN of the standby resource manager host. | rm-standby-host.acme.com | Optional |
To Enable High Availability, you must run several commands while starting cluster services.