This section provides general guidelines for cluster sizing and node requirements for effective use of the
|D s platform|
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.
Requirements for the
|D s node|
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:
- Data volume
- Number of concurrent jobs
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:
- If you are working with compressed or binary formats, you should use the expanded sizes for your data volume estimates.
- Some workloads are more compute- or memory-intensive and may increase the required number of nodes or capabilities of each node. These include:
- Scripts with complex steps such as joins (particularly those between large datasets) and sorts
- Lengthy scripts
- In high availability mode, the total number of connections across all nodes should meet the appropriate requirements in the above table. For each node, please divide the number of connections by the number of
D s item item nodes
Amazon Marketplace AMI
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
|D s product|
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
|D s proserv|
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.