Scaling services

A Red Hat® OpenShift® project administrator can adjust services by scaling them up to support high availability or to increase processing capacity.

Before you begin

Required role: To complete this task, you must be an OpenShift project administrator.

You can scale a service at any time after you install it.

About this task

Scaling a service changes its capacity by changing the number of replicas. For example, when you install and deploy a service on IBM® Cloud Pak for Data, the service is typically deployed with a single replica. If you increase the scale of the service, you increase the number of replicas. Increasing the number of replicas makes the service highly available. In some cases, it can also increase the processing capacity of the service.

Before you scale up a service, ensure that your cluster can support the additional workload. If necessary, contact your IBM Support representative.

The following table lists the services that you can scale. You set the scale of a services by specifying its assembly name in the scaling command. The table also lists any services that are affected when the specified service is scaled to medium.
Service name Scaling supported Assembly name Small Medium Large Notes
Cloud Pak for Data control plane Yes lite --config small --config medium --config large The default size is small.
Common core services Yes

Not applicable

Not applicable

Not applicable

Not applicable

 Scaled up by assemblies that rely on the common core services.
Analytics Engine Powered by Apache Spark Yes spark --config small --config medium

Not applicable

 The default size is small.
Cognos® Analytics No

Not applicable

Not applicable

Not applicable

Not applicable

  
Cognos Dashboards Yes cde 2 replicas

--config small

 4 replicas

--config medium

 8 replicas

--config large

 The default size is small.
Data Refinery Yes data-refinery 1 replica

--config small

 3 replicas

--config medium

Not applicable

 If you scale this service, the common core services are also scaled.
Data Virtualization Yes dv 1 replica

--config small

 2 replicas

--config medium

 3 replicas

--config large

 This service also supports an extra small configuration with 1 replica:

--config xsmall

The default size is xsmall.

The xsmall configuration sets smaller CPU limits for the head and worker nodes.

When replicas are scaled, the CPUs and memory are scaled accordingly.
DataStage® Enterprise Yes ds-ent 3 replicas, 2 CPUs, 6 Gi

--config small

 3 replicas, 4 CPUs, 12 Gi

--config medium

 4 replicas, 6 CPUs, 24 Gi

--config large

 When replicas are scaled, the CPUs and memory are scaled accordingly.
DataStage Enterprise Plus Yes ds 3 replicas, 2 CPUs, 6 Gi

--config small

 3 replicas, 4 CPUs, 12 Gi

--config medium

 4 replicas, 6 CPUs, 24 Gi

--config large

 When replicas are scaled, the CPUs and memory are scaled accordingly.
Db2® No

Not applicable

Not applicable

Not applicable

Not applicable

  
Db2 Big SQL Yes

Not applicable

Not applicable

Not applicable

Not applicable

 This service is scaled by using the Db2 Big SQL instance details page in the Cloud Pak for Data web interface.
Db2 Data Gate No

Not applicable

Not applicable

Not applicable

Not applicable

  
Db2 Data Management Console No

Not applicable

Not applicable

Not applicable

Not applicable

  
Db2 Event Store No

Not applicable

Not applicable

Not applicable

Not applicable

 Db2 Event Store requires at least three dedicated (labeled and tainted) nodes at all times. You can scale Db2 Event Store by changing the scale of the underlying virtual machines (VMs).
Db2 for z/OS® Connector No

Not applicable

Not applicable

Not applicable

Not applicable

  
Db2 Warehouse No

Not applicable

Not applicable

Not applicable

Not applicable

  
Decision Optimization Yes dods 1 replica, 2.5 CPUs, 4 GB

--config small

 2 replicas, 7 CPUs, 10 GB

--config medium

 3 replicas, 10 CPUs, 14 GB

--config large

 When replicas are scaled, the CPUs and memory are scaled accordingly.
Edge Analytics No

Not applicable

Not applicable

Not applicable

Not applicable

  
Execution Engine for Apache Hadoop No

Not applicable

Not applicable

Not applicable

Not applicable

  
Financial Services Workbench          
Guardium® External S-TAP® Yes ibm-estap-prod

Not applicable

Not applicable

Not applicable

 Use the oc scale command to set the number of replicas. The default is 2 replicas.
Jupyter Notebooks with Python 3.7 for GPU No

Not applicable

Not applicable

Not applicable

Not applicable

  
Jupyter Notebooks with R 3.6 No

Not applicable

Not applicable

Not applicable

Not applicable

  
Master Data Connect Yes mdc 2 replicas

--config small

 4 replicas

--config medium

 8 replicas

--config large

 Scaling also causes the CPU and memory limits to scale vertically on the supporting services that are shipped with the Master Data Connect service, including Apache Cassandra and Elasticsearch.
MongoDB No

Not applicable

Not applicable

Not applicable

Not applicable

 You can manually scale this service.
Open Data for Industries Yes ibm-osdu 2 replicas

--config small

 3 replicas

--config medium

 5 replicas

--config large

 This service also supports an extra small configuration with 1 replica:

--config xsmall
OpenPages® No

Not applicable

Not applicable

Not applicable

Not applicable

See Scaling OpenPages for an alternative method.
Planning Analytics Yes

This service scales resources, not replicas.

 pa --config pass-small
--config tm1-small		 --config pass-medium
--config tm1-medium		 --config pass-large
--config tm1-large		 Each resource can be scaled independently.
RStudio® Server with R 3.6 Yes rstudio

Not applicable

Not applicable

Not applicable

 If you scale this service, Watson™ Studio and common core services are also scaled.
SPSS® Modeler Yes spss-modeler If you scale this service, Watson Studio and common core services are also scaled.
Streams No

Not applicable

Not applicable

Not applicable

Not applicable

See Creating a Streams application resource template for an alternative method.
Streams Flows Yes streams-flows 1 replica

--config small

 3 replicas

--config medium

Not applicable

  
Virtual Data Pipeline No

Not applicable

Not applicable

Not applicable

Not applicable

  
Watson Assistant No

Not applicable

Not applicable

Not applicable

Not applicable

  
Watson Assistant for Voice Interaction No

Not applicable

Not applicable

Not applicable

Not applicable

  
Watson Discovery Yes

Not applicable

Not applicable

Not applicable

Not applicable

 See Custom scaling in Watson Discovery.
Watson Knowledge Catalog Yes wkc 1 replica for most services; 2 or 3 replicas for a few services

--config small

 A mix of 1, 2, or 3 replicas

--config medium

 A mix of 1, 3, or 4 replicas

--config large

 Legacy services always use 1 replica.
Watson Knowledge Studio Yes watson-ks 1 replica

--config small

 2 replicas

--config medium

Not applicable

  
Watson Language Translator          
Watson Machine Learning Yes wml 1 replica

--config small

 2 replicas

--config medium

Not applicable

 If you scale this service, the common core services are also scaled.
Watson Machine Learning Accelerator No

Not applicable

Not applicable

Not applicable

Not applicable

  
Watson OpenScale Yes aiopenscale  
Watson Speech to Text Yes watson-speech 1 replica

--config small

 2 replicas

--config medium

Not applicable

  
Watson Studio Yes wsl If you scale this service, the common core services are also scaled.
Watson Text to Speech Yes watson-speech 1 replica

--config small

 2 replicas

--config medium

Not applicable

  
Tip:

Procedure

  1. Change to the directory where you placed the Cloud Pak for Data command-line interface.
  2. Log in to your OpenShift cluster as a project administrator: 
    oc login OpenShift_URL:port
  3. To see a preview of what will happen when you scale up a service, run the following command: 
    ./cpd-cli scale \
--assembly Assembly_name \
--namespace Project \
--config Size \
--include-dependent-assemblies \
--dry-run

    If you installed multiple instances of the service in the same project, add the following line to your scale command after the --namespace flag:

    --instance Instance_name \
    Replace the following values:
    Variable Replace with
    Assembly_name Use the appropriate value for your service from the preceding table.
    Project The project where the service is installed.
    Size The size to scale to, for example, medium or large.
    Instance_name The service instance name that you want to scale up.
  4. Re-run the preceding command without the --dry-run flag.