Deployment of sample task fails in PCF

1
votes

spring-cloud-dataflow-server-2.0.1.RELEASE.jar

I am trying to deploy the sample task app on SCDF@PCF.

Deployment fails with the following Exception :

Shell side :

No Launcher found for the platform named 'default'.  Available platform names are []
org.springframework.cloud.dataflow.rest.client.DataFlowClientException: No Launcher found for the platform named 'default'.  Available platform names are []



SCDF Server side :

   2019-03-25T08:00:33.81-0500 [APP/PROC/WEB/0] OUT 2019-03-25 13:00:33.815 ERROR 19 --- [io-8080-exec-10] o.s.c.d.s.c.RestControllerAdvice         : Caught exception while handling a request
   2019-03-25T08:00:33.81-0500 [APP/PROC/WEB/0] OUT java.lang.IllegalStateException: No Launcher found for the platform named 'default'.  Available platform names are []
   2019-03-25T08:00:33.81-0500 [APP/PROC/WEB/0] OUT     at org.springframework.cloud.dataflow.server.service.impl.DefaultTaskExecutionService.findTaskLauncher(DefaultTaskExecutionService.java:199)
   2019-03-25T08:00:33.81-0500 [APP/PROC/WEB/0] OUT     at org.springframework.cloud.dataflow.server.service.impl.DefaultTaskExecutionService.executeTask(DefaultTaskExecutionService.java:151)
   2019-03-25T08:00:33.81-0500 [APP/PROC/WEB/0] OUT     at org.springframework.cloud.dataflow.server.service.impl.DefaultTaskExecutionService$$FastClassBySpringCGLIB$$422cda43.invoke(<generated>)

Any ideas ? Do I need to set a launcher ?

2
spring-cloud-dataflow

2 Answers

0
votes

It appears you may not have configured a platform for Tasks.

Starting from v2.0, SCDF provides the flexibility to configure multiple platform backends for Tasks, so you can choose from a list of platforms where you'd want to launch the Task. You can read more about the feature from the release highlights-blog.

If you haven't already configured the Task platform properties, please use the sample manifest.yml as a reference.

If you have set those properties and you still see this issue, feel free to share the manifest.yml - we can review for correctness. Of course, make sure to remove sensitive creds before sharing it.

0
votes

Just as complementary information:

I got the same error by launch on a Kubernetes platform (Openshift) and could resolve the problem by adding the following snippet in the application.yaml from dataflow-server:

spring:
  cloud:
    dataflow:
      task:
        platform:
          kubernetes:
            accounts:
              dev:
                namespace: devNamespace
                imagePullPolicy: Always
                entryPointStyle: exec
                limits:
                  cpu: 4
              qa:
                namespace: qaNamespace
                imagePullPolicy: IfNotPresent
                entryPointStyle: boot
                limits:
                  memory: 2048m

Reference: Documentation Dataflow