zentol commented on a change in pull request #10976: [FLINK-13978][build system] Add experimental support for building on Azure Pipelines URL: https://github.com/apache/flink/pull/10976#discussion_r378743130
########## File path: tools/azure-pipelines/jobs-template.yml ########## @@ -0,0 +1,137 @@ +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. + +parameters: + test_pool_definition: # defines the hardware pool for compilation and unit test execution. + e2e_pool_definion: # defines the hardware pool for end-to-end test execution + stage_name: # defines a unique identifier for all jobs in a stage (in case the jobs are added multiple times to a stage) + environment: # defines environment variables for downstream scripts + +jobs: +- job: compile_${{parameters.stage_name}} + condition: not(eq(variables['MODE'], 'e2e')) + pool: ${{parameters.test_pool_definition}} + container: flink-build-container + timeoutInMinutes: 240 + cancelTimeoutInMinutes: 1 + workspace: + clean: all # this cleans the entire workspace directory before running a new job + # It is necessary because the custom build machines are reused for tests. + # See also https://docs.microsoft.com/en-us/azure/devops/pipelines/process/phases?view=azure-devops&tabs=yaml#workspace + + steps: + # The cache task is persisting the .m2 directory between builds, so that + # we do not have to re-download all dependencies from maven central for + # each build. The hope is that downloading the cache is faster than + # all dependencies individually. + # In this configuration, we a hash over all committed (not generated) .pom files + # as a key for the build cache (CACHE_KEY). If we have a cache miss on the hash + # (usually because a pom file has changed), we'll fall back to a key without + # the pom files (CACHE_FALLBACK_KEY). + # Offical documentation of the Cache task: https://docs.microsoft.com/en-us/azure/devops/pipelines/caching/?view=azure-devops + - task: Cache@2 + inputs: + key: $(CACHE_KEY) + restoreKeys: $(CACHE_FALLBACK_KEY) + path: $(MAVEN_CACHE_FOLDER) + continueOnError: true # continue the build even if the cache fails. + displayName: Cache Maven local repo + + # Compile + - script: STAGE=compile ${{parameters.environment}} ./tools/azure_controller.sh compile + displayName: Build + + # upload artifacts for next stage + - task: PublishPipelineArtifact@1 + inputs: + path: $(CACHE_FLINK_DIR) + artifact: FlinkCompileCacheDir-${{parameters.stage_name}} + +- job: test_${{parameters.stage_name}} + dependsOn: compile_${{parameters.stage_name}} + condition: not(eq(variables['MODE'], 'e2e')) + pool: ${{parameters.test_pool_definition}} + container: flink-build-container + timeoutInMinutes: 240 + cancelTimeoutInMinutes: 1 + workspace: + clean: all + strategy: + matrix: + core: + module: core + python: + module: python + libraries: + module: libraries + blink_planner: + module: blink_planner + connectors: + module: connectors + kafka_gelly: + module: kafka/gelly + tests: + module: tests + legacy_scheduler_core: + module: legacy_scheduler_core + legacy_scheduler_tests: + module: legacy_scheduler_tests + misc: + module: misc + steps: + + # download artifacts + - task: DownloadPipelineArtifact@2 + inputs: + path: $(CACHE_FLINK_DIR) + artifact: FlinkCompileCacheDir-${{parameters.stage_name}} + + # recreate "build-target" symlink for python tests + - script: | + ls -lisah $(CACHE_FLINK_DIR) + ls -lisah . + ln -snf $(CACHE_FLINK_DIR)/flink-dist/target/flink-*-SNAPSHOT-bin/flink-*-SNAPSHOT $(CACHE_FLINK_DIR)/build-target + displayName: Recreate 'build-target' symlink + # Test + - script: STAGE=test ${{parameters.environment}} ./tools/azure_controller.sh $(module) + displayName: Test - $(module) + + - task: PublishTestResults@2 + inputs: + testResultsFormat: 'JUnit' + + +- job: e2e_${{parameters.stage_name}} + condition: eq(variables['MODE'], 'e2e') + # We are not running this job on a container, but in a VM. Review comment: But the pre-commit-tests are still being run in docker, no? azure-controller.sh#L191: ` # Some tests in the "run-pre-commit-tests.sh" collection launch Docker containers. # Since the regular build is executed in Docker (on Azure), we'll be launching those # containers outside of the current container (on the host, alongside the build&test container). # Some of these containers mount a path. Currently, these scripts mount relative to the build container, # thus this path is not available on the host (where the test container is launched). # # Here, we figure out the path on the host machine, and set it. #` If running end-to-end tests in a docker container is troublesome, shouldn't the pre-commit-tests also be executed separately? As it stands we have to handle both setups. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services