Volume claim is generated with access mode ReadWriteMany for default storage class

[blublinsky]

/kind bug

Here is a simple KFP application:

from kfp.components import InputPath, InputTextFile, OutputPath, OutputTextFile
from kfp.components import func_to_container_op
from kfp_tekton.compiler import TektonCompiler

# Writing many numbers
@func_to_container_op
def write_numbers(numbers_path: OutputPath(str), start: int = 0, count: int = 10):
    with open(numbers_path, 'w') as writer:
        for i in range(start, count):
            writer.write(str(i) + '\n')


# Reading and summing many numbers
@func_to_container_op
def sum_numbers(numbers_path: InputPath(str)) -> int:
    sum = 0
    with open(numbers_path, 'r') as reader:
        for line in reader:
            sum = sum + int(line)
    return sum

# Printing
@func_to_container_op
def print_text(text_path: InputPath()): # The "text" input is untyped so that any data can be printed
    '''Print text'''
    with open(text_path, 'r') as reader:
        for line in reader:
            print(line, end = '')

# Pipeline to sum 100000 numbers
def sum_pipeline(count: 'Integer' = 100000):
    numbers_task = write_numbers(count=count)
    print_text(numbers_task.output)

    sum_task = sum_numbers(numbers_task.outputs['numbers'])
    print_text(sum_task.output)

yamlFile = 'sum_pipeline.yaml'
TektonCompiler().compile(sum_pipeline, yamlFile)

The execution requires a PVC for execution.

But if you look at generated yaml file, you will see

..................
  workspaces:
  - name: sum-pipeline
    volumeClaimTemplate:
      spec:
        accessModes: [ReadWriteMany]
        resources:
          requests: {storage: 2Gi}

Which requires read write many access mode for a default storage class, which fails on PVC binding

What do I expext

..................
  workspaces:
  - name: sum-pipeline
    volumeClaimTemplate:
      spec:
        accessModes: [ReadWriteOnce]
        resources:
          requests: {storage: 2Gi}

Which actually works

Environment:

• Python Version 3.7
• SDK Version: 0.8.1
• Tekton Version (use tkn version): v0.24
• Kubernetes Version (use kubectl version): v1.19.0+d856161
• OS (e.g. from /etc/os-release):

[Tomcli]

You can configure the workspace mounting strategy using these env variables https://github.com/kubeflow/kfp-tekton/tree/master/sdk#big-data-passing-workspace-configuration

[blublinsky]

Thanks @Tomcli, it works great

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[Tomcli]

/close fixed

[google-oss-prow[bot]]

@Tomcli: Closing this issue.

In response to this:

/close fixed

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.