After the IBM Streams service
is installed, you must provision a Streams service instance before you can
process streaming data. Provisioning the Streams service instance creates and
starts a Streams instance and
a build service. You can provision multiple service instances for Streams.
Before you begin
Gather the following information that you must enter when you
provision a service instance. For more information, see Preparing to install and provision the Streams service.
- The external libraries storage class or persistent volume claim.
- The instance state cache storage class or persistent volume claim.
- The build service storage class or persistent volume claim.
- (Optional) The job checkpoint repository.
- (Optional) The builder service storage class or persistent volume claim.
About this task
To complete this task, you must have the Administrator and
Create service instances permissions. By default, an administrator has these
permissions.
You can provision multiple service instances for
Streams. You might want multiple
instances for the following reasons:
- Isolating user access
- Isolating applications or jobs
- Maintaining a development and a production environment
Procedure
To provision a Streams service instance:
-
From the IBM Cloud Pak for Data web user interface
navigation menu, click .
-
From the list of services, select the IBM Streams tile.
-
Depending on whether there are already Streams instances, click
Provision instance or select New instance from the
tile options menu.
-
Configure the Streams
service instance by setting the required and optional parameters.
-
To enable the Streams
console, select Enable the Streams Console on the Streams instance
services step.
-
To access the streamtool command from the operations pod, select
Enable the operations pod on the Streams instance
services step.
-
To enable application image builds for the edge, select Enable image builds for the
edge on the Streams instance service step.
-
To use Streams Flows, a builder
persistent volume claim must be created. If you choose to dynamically provision persistent storage
for Streams, the builder
persistent volume claim is created by default. Otherwise, you must specify a value for
Builder persistent volume claim.
-
Click Next.
-
Ensure that the summary is correct and click Provision.
Important:
The amount of time required to provision the Streams service instance depends on
many factors, such as the number of external libraries that are created and the performance of the
file systems that are used. The process can take 10 - 60 minutes. The provisioning statuses can help
you determine if the instance is being provisioned or if an error occurred that prevented the
provisioning from completing successfully.
For example, a successfully provisioned service instance has an overall status of
running, a Streams instance status of
Running, and a build status of Ready. When the
provisioning is in progress, the service instance might have an overall service instance status of
pending, a Streams instance status of
Running, and a build status of Installing. This
scenario indicates that the build service is working on installing an external library and that the
provisioning is proceeding as expected.
For more information about the provisioning statuses, see Service statuses (Streams). For information about viewing the service
status, see Managing Streams. For information about troubleshooting
provisioning issues, see The service instance provisioning fails (Streams).
What to do next
- If you want other users to be able to manage or use this service instance, you as the service
instance creator and administrator must grant them permission. Administrator access is required in
order to manage the service instance. User access is required in order to create, build, or submit a
Streams application in the
IBM Streams instance.
- You can use streams flows to develop and deploy streaming applications by using the Streams Flows app. For information about
installing Streams Flows, see Installing
Streams Flows.
- You can develop streams-processing applications by using notebooks. For more information, see
Getting started with Python notebooks (Streams).
- You can develop streaming applications that use Streams Processing Language (SPL) and
deploy the applications to Cloud Pak for Data. For
more information, see Developing Streams Processing Language applications.