Fix checkpoint flush retry and WAL reset handling

[PhantomInTheWire]

Summary

• Retry flush() checkpoints when a concurrent write invalidates staged checkpoint state instead of returning success after discard.
• Preserve committed checkpoint data when WAL reset fails, leaving stale WAL for recovery instead of rolling back data/manifest files.
• Add regression coverage for stale-WAL reset failure and deterministic concurrent-write flush retry behavior.
• Simplify the branch changes to avoid retry loops, sleeps, and unnecessary WAL-reset branching.

Testing

• cargo fmt --check
• git diff --check
• cargo test -p garuda-engine --test test_collection_recovery_crash_semantics
• cargo test -p garuda-engine

  

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: e1dd460b-7a03-4c67-b4f2-0ab8a8881cda

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch cursor/fix-checkpoint-flush-wal-82b1

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}