[hotfix] disk space and file descriptor leaks in ChangelogStreamHandleReaderWithCache

[Manishnemade12]

[hotfix] disk space and file descriptor leaks in ChangelogStreamHandleReaderWithCache

What is the purpose of the change

This pull request fixes severe disk space and file descriptor leaks in ChangelogStreamHandleReaderWithCache that occur during filesystem or network errors. This is a hotfix because the leaks can cause TaskManager crashes due to exhausted disk space or file descriptors and should be applied to affected stable branches as appropriate.

Previously, if a network cluster error disrupted DFS stream copying in downloadToCacheFile, the partially transferred temporary cache file was permanently left on disk, eventually causing disk full crashes on TaskManagers. Furthermore, if openAndSeek threw an IOException during channel positioning, the instantiated FileInputStream was leaked without being closed, exhausting OS file handles.

Brief change log

• Explicitly invoke file.delete() on the target temporary block when an IOException is encountered during DFS buffer copying in downloadToCacheFile, ensuring partial temporary files are removed on failures.
• Wrap the fin.getChannel().position(offset) call in openAndSeek with a try-catch that guarantees IOUtils.closeQuietly(fin) is called if positioning fails, preventing leaked FileInputStream instances.

Verifying this change

This change is already covered by existing tests, such as the Changelog State Backend recovery tests that implicitly exercise filesystem operations and DFS stream caching logic.

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: yes (Improves resilience of Checkpointing recovery on TaskManagers)
• 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

[flinkbot]

CI report:

• 78d5745 Azure: SUCCESS

Bot commands

The @flinkbot bot supports the following commands:

• @flinkbot run azure re-run the last Azure build

[spuru9]

@Manishnemade12 Can you add [hotfix] to the PR and update the PR description as per the new guidelines.

[Manishnemade12]

@Manishnemade12 Can you add [hotfix] to the PR and update the PR description as per the new guidelines.

@spuru9 i did changes . can you please review it now

[spuru9]

You would need to run mvn spotless:apply

[Manishnemade12]

@spuru9 i addressed all comment , can you please took once look there

build is also fine