h2o_init_multinode: Starts an h2o cluster on multiple nodes
Description
Can be used on a SLURM system or any cluster where h2o can take advantage
of multiple computers.
Usage
h2o_init_multinode(
h2o_jar,
node_list = Sys.getenv("SLURM_NODELIST"),
memory_per_node = "8g",
cluster_name = "h2o_r",
verbose = T,
...
)
Arguments
h2o_jar
Path to h2o.jar, e.g. "~/software/h2o-latest/h2o.jar"
node_list
Comma-separated list of nodes to use, defaults to
the value of SLURM_NODELIST. Assumed to be hostnames, not IPs.
memory_per_node
Memory per node. Default 8g.
cluster_name
Name of the h2o cluster. Default "h2o_r".
verbose
If T display extra information during execution.
...
Remaining arguments passed to h2o.init()
Examples
Run this code# NOT RUN {
cl = h2o_init_multinode("~/software/h2o-latest/h2o.jar")
h2o.shutdown()
# }
# NOT RUN {
# }
Run the code above in your browser using DataLab