Nodes setup types
On Cloud Pak for Data System 2.0.X, the Netezza Performance Server host runs in the Red Hat OpenShift container orchestration environment along with Cloud Pak for Data.
Learn about the Red Hat OpenShift nodes setup options that you can choose from when you are installing Netezza Performance Server 11.2.2.X on Cloud Pak for Data System 2.0.X.
- Dedicated nodes setup
-
The dedicated nodes setup provides an isolated set of nodes for the Netezza Performance Server host and other services (backup and restore, replication) to use.
- Advantages
-
- Cloud Pak for Data services and other Netezza Performance Server tenants are forbidden from running on these nodes, providing a dedicated environment for Netezza Performance Server.
- The nodes are put in their own dedicated machine configuration pool
(
nps-dedicated-<namespace>). Node updates that are applied by the Cloud Pak for Data System upgrade, which trigger a rolling restart, are not applied to these nodes until a downtime for this instance can be scheduled. - Different tenants have different
nps-dedicated<namespace>machine config pools. Each instance can do a Cloud Pak for Data System node upgrade at different times.
- Disadvantages
-
- Each tenant needs to set aside at least two nodes for HA purposes.
- Tenants cannot share a node for HA.
- Cloud Pak for Data services cannot use these nodes.
- Shared nodes setup
-
With the shared nodes setup, Cloud Pak for Data and other Netezza Performance Server tenants might share a common set of nodes.
- Advantages
-
- When in a multi-tenant environment, different tenants can share a common set of nodes. For two tenants, you need only three nodes instead of four for to install Netezza Performance Server because both instances can share a node for HA.
- Cloud Pak for Data services prefer to not run on Netezza Performance Server nodes, but the services might still use the nodes if the overall system is constrained for resources.
- Disadvantages
-
- All Netezza Performance Server tenants that are set
up as shared share nodes that are put into the
nps-sharedmachine configuration pool. When a Cloud Pak for Data System update applies changes, which require node restart, all Netezza Performance Server instances in the shared setup need to schedule downtime at the same time. - In a resource-constrained environment, the nodes might become saturated with Cloud Pak for Data pods. The Netezza Performance Server host pod has priority over these pods, and during a failover scenario evicts lower priority pods so that the host can run. This process might take some time as compared to having a dedicated node with no resource constraints.
- All Netezza Performance Server tenants that are set
up as shared share nodes that are put into the
Only Netezza Performance Server pods might run on the dedicated setup so fewer resources are available for Cloud Pak for Data.
The Cloud Pak for Data base can likely support only two tenants that run in the shared setup, since the Cloud Pak for Data System platform also requires Cloud Pak for Data resources.
If you use Cloud Pak for Data, you might want to expand to a greater number of nodes to support multiple Netezza Performance Server tenants alongside Cloud Pak for Data.