When an instance of H2O is started from the command line, users generally call a java command similar to “java -Xmx1g -jar h2o.jar ” Users can customize their running instance of H2O by changing options in this java command. Each of the options has a default.
Help in the terminal:
Users can access help while working in the terminal by calling java -jar h2o.jar -help in the working directory where their H2O jar is located.
Memory Allocation
In java [-Xmx<size>] -jar h2o.jar [options] users specify the amount of memory allocated to a particular node by specifying a number in place of the indicator <size>. If your data set is large, give H2O more memory (for example, -Xmx4g gives H2O four gigabytes of memory). For best performance, Xmx should be 4x the size of your data, but never more than the total amount of memory on the machine where H2O is running.
Where to Place Options
There are two places in the java command where options can be specified. In the call options should be inserted as follows: java -Xmx<size> [ADDITIONAL JVM OPTIONS GO HERE] -jar h2o.jar [H:sub:`2`O OPTIONS GO HERE]
The common options are listed below; however, H2O is designed to not require many flags. Advanced and curious users may find http://www.oracle.com/technetwork/java/javase/tech/vmoptions-jsp-140102.html helpful for defining and describing other common java command options.
-version
Print java version info and exit.
-Xmx[totalheapsize]
Sets the total heap size for the node of H2O.
-flatfile <FlatFileName>
Configuration flat file explicitly listing H2O cloud node members.
-h | -help
Print this help.
-ice_root <fileSystemPath>
The directory where H2O spills temporary data to disk. (The default is ‘/tmp/h2o-User’.)
-ip <ipAddressOfNode>
IP address of this node.
-name <h2oCloudName>
Cloud name used for discovery of other nodes. Nodes with the same cloud name will form an H2O cloud (also known as an H2O cluster).
-network <IPv4network1Specification>[,<IPv4network2Specification> …]
The IP address discovery code will bind to the first interface that matches one of the networks in the comma-separated list. Use instead of -ip when a broad range of addresses is legal. (Example network specification: ‘10.1.2.0/24’ allows 256 legal possibilities.)
-nthreads <number>
Maximum number of typical worker threads. Think of this as the maximum number of CPUs H2O will use. This value is per java instance. Specifying too small a value may result in deadlocks. Never specify fewer than 4. The default value is 99.
-port <port>
Port number for this node (note: port+1 is also used). (The default port is 54321.)
New H2O nodes join together to form a cloud at startup time. Once a cloud is given work to perform, it locks out new members from joining. H2O works best on distributed multinode clusters when the clusters are similar in configuration, and allocated equal amounts of memory.
Examples:
- Start an H2O node with 4GB of memory and a default cloud name:
- $ java -Xmx4g -jar h2o.jar
- Start an H2O node with 6GB of memory and a specified cloud name:
- $ java -Xmx6g -jar h2o.jar -name TomsCloud
- Start an H2O cloud with three 2GB nodes and a specified cloud name:
- $ java -Xmx2g -jar h2o.jar -name TomsCloud $ java -Xmx2g -jar h2o.jar -name TomsCloud $ java -Xmx2g -jar h2o.jar -name TomsCloud