[Failing Test]: XVR test flaky build python SDK container imager: "/target/launcher/linux_amd64/boot" not found

[Abacn]

What happened?

Saw this a couple of times before.

e.g. https://ci-beam.apache.org/job/beam_PostCommit_XVR_Flink_PR/299/consoleFull

12:53:21 #2 [internal] load .dockerignore
12:53:21 #2 transferring context: 2B done
12:53:21 #2 DONE 0.5s
12:53:21 
12:53:21 #3 [internal] load metadata for docker.io/library/python:3.8-bullseye
12:53:21 #3 DONE 0.5s
12:53:21 
12:53:21 #4 [beam 1/5] FROM docker.io/library/python:3.8-bullseye@sha256:5598a0361e5a1e6fc0bcf942208ac47bf17b1865ea7767c2f1e6d572f7917736
12:53:21 #4 CACHED
12:53:21 
12:53:21 #5 [internal] load build context
12:53:21 #5 transferring context: 2.49MB 0.2s done
12:53:21 #5 DONE 0.3s
12:53:21 
12:53:21 #6 [beam 3/5] COPY target/apache-beam.tar.gz /opt/apache/beam/tars/
12:53:21 #6 CACHED
12:53:21 
12:53:21 #7 [beam 4/5] COPY target/launcher/linux_amd64/boot target/LICENSE target/NOTICE target/LICENSE.python /opt/apache/beam/
12:53:21 #7 ERROR: "/target/launcher/linux_amd64/boot" not found: not found
12:53:21 
12:53:21 #8 [beam 2/5] COPY target/base_image_requirements.txt /tmp/base_image_requirements.txt
12:53:21 #8 CANCELED
12:53:21 ------
12:53:21  > [beam 4/5] COPY target/launcher/linux_amd64/boot target/LICENSE target/NOTICE target/LICENSE.python /opt/apache/beam/:
12:53:21 ------
12:53:21 ERROR: failed to solve: failed to compute cache key: "/target/launcher/linux_amd64/boot" not found: not found

Issue Failure

Failure: Test is flaky

Issue Priority

Priority: 2 (backlog / disabled test but we think the product is healthy)

Issue Components

• Component: Python SDK
• Component: Java SDK
• Component: Go SDK
• Component: Typescript SDK
• Component: IO connector
• Component: Beam YAML
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner

[Abacn]

The failed test has

Task :sdks:python:container:py38:copyLauncherDependencies NO-SOURCE

which means this copy job had no source, i.e. :sdks:python:container:goBuild not producing the launcher or it gets immediately deleted?

[Abacn]

This is likely a gradle bug, see #23785 (comment)

It is now flaky on both Jenkins and GitHub Actions. It affects a couple of XVR jobs Flink,Spark,etc

A reproduce: https://ci-beam.apache.org/job/beam_PostCommit_XVR_Flink_PR/324/console

• At gradle configuration time, configurations.pythonHarnessLauncher is indeed set:

13:04:47 > Configure project :sdks:python:container:py38
13:04:47 print----- /home/jenkins/jenkins-slave/workspace/beam_PostCommit_XVR_Flink_PR/src/sdks/python/container/build/target/launcher
13:04:48 System Go installation: /snap/bin/go is go version go1.16.15 linux/amd64; Preparing to use /home/jenkins/go/bin/go1.21.3
13:04:48 go1.21.3: already downloaded in /home/jenkins/sdk/go1.21.3
13:04:48 GOCMD=/home/jenkins/go/bin/go1.21.3
13:04:49 print+++++ /home/jenkins/jenkins-slave/workspace/beam_PostCommit_XVR_Flink_PR/src/sdks/python/container/build/target/launcher

• Then, :sdks:python:container:goBuild indeed produced the binary:

13:05:39 > Task :sdks:python:container:goBuild
13:05:39 print:::::  /home/jenkins/jenkins-slave/workspace/beam_PostCommit_XVR_Flink_PR/src/sdks/python/container/build/target/launcher/linux_amd64/boot name:::: boot

However, sdks:python:container:py38:copyLauncherDependencies skipped as NO-SOURCE

13:05:39 > Task :sdks:python:container:py38:copyLauncherDependencies NO-SOURCE

which caused binary not copied into the folder for docker prepare

[Abacn]

Possibly related to gradle/gradle#24368