flink icon indicating copy to clipboard operation
flink copied to clipboard

[FLINK-29797][flink-yarn] Fix fs.default-scheme will accidentally cau…

Open slankka opened this issue 10 months ago • 1 comments

…se temporary flinkConfigFile treated as remote file. also fix [FLINK-33424]

What is the purpose of the change

If user set fs.default-scheme to hdfs://namenode or s3://any, it will cause /tmp/flink-confxxxx FileNotFoundException.

Fix FLINK-29797

also fix FLINK-33424

Brief change log

It is obvious that FileSystem will create a DistributedFileSystem while fs.default-scheme NOT set to file://.

It is obvious that temporary flink-conf.yaml or config.yaml file are created as LocalResource during submission to yarn.

Before change

There is no big diffierence between "job.graph" and flink conf file for registerSingleLocalResource except that tmpJobGraphFile.toURI() and tmpConfigurationFile.getAbsolutePath()

After change

tmpConfigurationFile.toURI() will return correct path with file:// scheme, so registerSingleLocalResource will not treate it as remote files.

Please be aware of that registerSingleLocalResource in file flink/flink-yarn/src/main/java/org/apache/flink/yarn/YarnClusterDescriptor.java at master · apache/flink (github.com) has two similar code:

uploading tmpConfigurationFile and tmpJobGraphFile using different way to get path.

//Line: 1104
fileUploader.registerSingleLocalResource(
                    flinkConfigFileName,
                    new Path(tmpConfigurationFile.getAbsolutePath()),  //originally tmpConfigurationFile.toUri()
                    "",
                    LocalResourceType.FILE,
                    true,
                    true); 

[FLINK-18362][yarn] Fix mistakenly merged commit 0e10fd5b8ee0 · apache/flink@a0227e2 (github.com)

https://github.com/apache/flink/commit/a0227e20430ee9eaff59464023de2385378f71ea

//Line: 1071              
fileUploader.registerSingleLocalResource(
                        jobGraphFilename,
                        new Path(tmpJobGraphFile.toURI()),
                        "",
                        LocalResourceType.FILE,
                        true,
                        false);

Verifying this change

This change is a trivial rework / code cleanup without any test coverage.

Does this pull request potentially affect one of the following parts:

  • Dependencies (does it add or upgrade a dependency): ( no)
  • The public API, i.e., is any changed class annotated with @Public(Evolving): ( no)
  • The serializers: ( no )
  • The runtime per-record code paths (performance sensitive): ( no )
  • Anything that affects deployment or recovery: JobManager (and its components), Checkpointing, Kubernetes/Yarn, ZooKeeper: (yarn)
  • The S3 file system connector: ( no )

Documentation

  • Does this pull request introduce a new feature? (no)
  • If yes, how is the feature documented? (not applicable)

slankka avatar Apr 27 '24 11:04 slankka

CI report:

  • b897027bc10b1311e55343f114c77675d74aaeb0 Azure: SUCCESS
Bot commands The @flinkbot bot supports the following commands:
  • @flinkbot run azure re-run the last Azure build

flinkbot avatar Apr 27 '24 11:04 flinkbot

@flinkbot run azure

ferenc-csaky avatar Oct 08 '24 18:10 ferenc-csaky