This section provides general guidelines for cluster sizing and node requirements for effective use of the .
NOTE: These guidelines are rough estimates of what should provide satisfactory performance. You should review particulars of the variables listed below in detail prior to making recommendations or purchasing decisions.
See System Requirements.
All compute nodes on the cluster (Hadoop NodeManager nodes) should have identical capabilities. Avoid mixing and matching nodes of different capabilities.
Primary variables affecting cluster size:
In the following table, you can review the recommended number of worker nodes in the cluster based on the data volume and the number of concurrent jobs. Table data assumes that each compute node has 16 compute cores (2 x 8 cores), 128GB of RAM and 8TB of disk, with nodes connected via 10 gigabit Ethernet (GbE).
|Data Volume \ Number of concurrent jobs||1||5||10||25|
|1 GB or less||1||1||1||2|
|1000 GB (1 TB)||20||100||200||500|
Additional variables affecting cluster size:
Amazon Marketplace installations support a limited range of installation options for the AMI. For more information, see the install guide available through the Marketplace for .
NOTE: The sizing guidelines listed for Enterprise Hadoop above provide a good estimate for sizing capacity and upper bounds for EMR-based cluster scaling.
For additional details on sizing your EMR cluster, please contact .
Microsoft Azure installations support a limited range of installation options, based on the type of cluster integration.
Please use the Enterprise Hadoop guidelines listed previously.
For more information on this integration, see Configure for HDInsight in the Configuration Guide.
Please review the Enterprise Hadoop guidelines with .
For more information on this integration, see Configure for Azure Databricks in the Configuration Guide.