System Administration
Also available as:
PDF

Region Count and Size

In general, an HBase cluster runs smoother with fewer regions. However, administrators cannot directly configure the number of regions for a RegionServer. However, administrators can indirectly increase the number of regions in the following ways: In addition, administrators can indirectly affect the number of regions for a RegionServer in the following ways:

  • Increase the size of the memstore for a RegionServer

  • Increase the size of a region

In addition, administrators can increase the number of regions for a RegionServer by pre-splitting large regions to spread data and the request load across the cluster. HBase allows administrators to individually configure each HBase table, which is useful when tables have different workloads and use cases. Most region settings can be set on a per-table basis with HTableDescriptor class, as well as the HBase CLI. These methods override the properties in the hbase-site.xml configuration file. For further information, see Configure Compactions.

[Note]Note

The HDFS replication factor affects only disk usage and should not be considered when planning the size of regions. The other factors described in the table above are applicable.