address comment

[liunyl]

Here are some reminders before you submit the pull request

• Add tests for the change
• Document changes
• Reference the link of issue using fixes eloqdb/tx_service#issue_id
• Reference the link of RFC if exists
• Pass ./mtr --suite=mono_main,mono_multi,mono_basic

Summary by CodeRabbit

• Refactor
  • Optimized internal memory accounting for data synchronization operations with improved precision in buffer allocation calculations across different system configurations.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

Ubuntu seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[coderabbitai]

Caution

Review failed

The pull request is closed.

Walkthrough

This pull request refactors memory accounting for flush data size in local_cc_shards.cpp by introducing conditional compilation logic. When WITH_JEMALLOC is defined, it uses sizeof(FlushRecord) for accounting; otherwise, it uses mi_malloc_usable_size() on vector buffers to measure actual allocated memory for DataSyncVec, ArchiveVec, and MoveBaseIdxVec.

Changes

Cohort / File(s)	Summary
Memory accounting refactor tx_service/src/cc/local_cc_shards.cpp	Introduces conditional memory sizing based on WITH_JEMALLOC flag. When jemalloc is enabled, uses fixed sizeof(FlushRecord) for all three vectors (DataSyncVec, ArchiveVec, MoveBaseIdxVec). When disabled, measures actual usable buffer sizes via mi_malloc_usable_size(). Applied consistently in both DataSyncForRangePartition and DataSyncForHashPartition code sections.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• #323: Modifies the same file and memory accounting logic for FlushRecord/vector buffers, exploring sizeof versus malloc_usable_size trade-offs.

Suggested reviewers

• xiexiaoy
• MrGuin

Poem

🐰 A rabbit hops through memory lanes so bright,
Choosing paths with jemalloc's might,
Or mimalloc's malloc_usable_size—
Each buffer measured, memory precise,
Three vectors dance in perfect sync,
No more guessing, just the facts we think! 🎪

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch index_mem

---

📜 Recent review details

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between c5f9174 and 2e21758.

📒 Files selected for processing (1)

• tx_service/src/cc/local_cc_shards.cpp

---

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.}