org.apache.spark.h2o.backends.external
Timeout in milliseconds for watchdog client connection.
Timeout in milliseconds for watchdog client connection. If client is not connected to the external cluster in the given time, the cluster is killed
Timeout in milliseconds specifying how often the check for connected watchdog client is done
Driver IP address in case of auto mode in external cluster backend
Path to h2o driver
HDFS dir for external h2o nodes when starting in auto mode
Health check interval for external H2O nodes
Path to a file used as cluster notification file
If true, the client will try to kill the cluster and then itself in case some nodes in the cluster report unhealthy status
How often check the healthy status for the decision whether to kill the cloud or not.
ip:port of arbitrary h2o node to identify external h2o cluster
If this option is set to "auto" then h2o external cluster will be automatically started using the provided h2o driver on yarn, otherwise it is expected that the cluster will be started by the user
Timeout in seconds for starting h2o external cluster
Yarn queue on which external cluster should be started
Number of memory assigned to each external h2o node when starting in auto mode
Number of nodes to wait for when connecting to external H2O cluster in manual mode.
Number of nodes to wait for when connecting to external H2O cluster in manual mode. In auto mode, number of nodes to be started.
Kerberos principal
Kerberos principal
Timeout in seconds for confirmation of read operation ( h2o frame => spark frame) on external cluster.
Impersonated Hadoop user
Timeout in seconds for confirmation of write operation ( spark frame => h2o frame) on external cluster.