Deploying Hue with a ResourceManager HA Cluster
Add the HA configuration sub-section at the end of the yarn_clusters
section of the hue_ini
configuration file.
Add the HA sub-section
If it does not exist, you can add a [[[ha]]]
configuration
sub-section to end of the [[yarn_clusters]]
section of the
hue.ini
configuration file.
To do this, specify the following:
-
resourcemanager_api_url
Host on which you are running the failover Resource Manager.
-
logical_name
Logical name of the Resource Manager.
-
submit_to
Specify if Hue should use the cluster. Specify
true
to use the cluster, orfalse
to not use the cluster.
If you ensure that the cluster section names are unique, you can add more than one
Resource Manager failover configuration, for example, [[[ha2]]]
. By doing
so, if the standby Resource Manager also fails over, the next available RM instance is
tried.
Example:
[[yarn_clusters]] [[[default]]] # Whether to submit jobs to this cluster submit_to=true ## security_enabled=false # Resource Manager logical name (required for HA) logical_name=rm1 # URL of the ResourceManager webapp address (yarn.resourcemanager.webapp.address) resourcemanager_api_url=http://cl2-node02.local:8088 # URL of Yarn RPC adress (yarn.resourcemanager.address) resourcemanager_rpc_url=http://cl2-node02.local:8050 # URL of the ProxyServer API proxy_api_url=http://cl2-node02.local:8088 # URL of the HistoryServer API history_server_api_url=http://cl2-node02.local:19888 # URL of the NodeManager API node_manager_api_url=http://localhost:8042 # HA support by specifying multiple clusters # e.g. [[[ha]]] # Enter the host on which you are running the failover Resource Manager resourcemanager_api_url=http://cl2-node01.local:8088 history_server_api_url=http://cl2-node02.local:19888 proxy_api_url=http://cl2-node01.local:8088 resourcemanager_rpc_url=http://cl2-node01.local:8050 logical_name=rm2 submit_to=True Note only one ha section is necessary, since Hue still uses the default section.