Resource groups

Resource groups are logical groups of hosts that help organize a heterogeneous resource pool.

Understanding resource groups
Resource groups provide a simple way of organizing and grouping resources (hosts). Instead of creating policies for individual resources, you can create and apply them to an entire group. Groups can be made of resources that are explicitly listed by name or those that satisfy a specific static requirement in terms of operating system, memory, swap space, CPU factor, and so on.
Creating initial resource groups
Setting up useful resource groups is essential to making full and efficient use of cluster capabilities. Follow these steps to create resource groups that set aside specific hosts for management duties and divide up the remainder of your hosts that are based on maximum memory.
Creating resource groups
You can create a resource group by host name (static resource group using a list of hosts) or by resource requirement (dynamic resource group using static requirements).
Best practices for resource group configuration
When you create a Spark instance group, you can configure your own resource settings to avoid resource competition and unwanted resource reclaims.
Modifying resource groups
Modify resource groups to refine resource allocations used in your resource plan. You may also want to add or remove hosts from resource groups if you have recently added or removed hosts in your cluster.
Viewing hosts in a resource group
View hosts belonging to a resource group using the cluster management console.
Removing resource groups
Remove a resource group if it no longer meets your needs or is not being used. Hosts must not belong to multiple groups, so when you create new groups, ensure that you modify or delete the old groups.
Excluding resource groups from the host view
It is sometimes easier to examine the workload status of a host when you can exclude one or more resource groups from the host view.
Ranking resource groups in order of preference
When a Spark instance group can use two or more resource pools and prefers to receive slots from some resource pools over others, rank the resource groups for that consumer in order of preference. With this configuration, the Spark instance group gets its planned share of resources from each resource group, starting from the one with the highest rank.

Parent topic: Resource sharing and distribution