By default, MapR automatically uses all available network interface cards (NICs) on each node in a network. However, in some scenarios you might want MapR to use a restricted subnet of NICs. For example, if you use multiple NICs of mixed speeds (such as 1GbE and 10GbE) on each node, you might want to separate them into two subnets. That way, you can use the faster NICs for MapR and the slower NICs for other functions.
To illustrate this arrangement, the following diagram shows three nodes, each with a 1GbE NIC (eth0) and a 10GbE NIC (eth1). All the 1GbE NICs are networked together and connected to Network A. Likewise, all the 10GbE NICs are networked together and connected to Network B, where peak performance is required.
In the example above, the faster eth1 NICs are connected together as part of a subnet written as 10.10.0.0/16 in CIDR notation. To make sure that MapR only uses the subnet with the faster NICs, set the
MAPR_SUBNETS environment variable in
$MAPR_HOME/conf/env.sh as shown:
When you restrict MapR to certain subnets, only the designated subnets have full access to MapR clients. The other NICs on the network can only access these MapR clients:
- NFS client
- JobTracker client
- MCS client
Users who access the network through NICs that are outside the designated subnets:
- Cannot execute Hadoop commands from the command line
- Cannot launch MapReduce jobs