BUG: Connection leak in db_update_cluster() causes crashes in transaction-based operations

[yugalkaushik]

Is there an existing issue for this?

• I have searched the existing issues

What happened?

The db_update_cluster() function in face_clusters.py unconditionally closes database connections in its finally block, even when the connection was passed from external code. This causes crashes in transaction-based operations where the caller manages the connection lifecycle.

Impact:

• Crashes in transaction-based operations
• Externally-managed connections are prematurely closed
• Database integrity issues when operations fail mid-transaction

Record

• I agree to follow this project's Code of Conduct

[coderabbitai]

📝 CodeRabbit Plan Mode

Generate an implementation plan and prompts that you can use with your favorite coding agent.

• Create Plan

Examples

• Example 1
• Example 2

---

🔗 Related PRs

#547 - Fix: Critical database connection leaks in all database functions [merged]
#570 - Fix: Ensure atomic DB operations and consistency in global reclustering (#562) [merged]

👤 Suggested Assignees

• Aditya30ag
• Hemil36

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord or schedule a call!

[V4Vikaskumar]

Hi 👋

I’d like to work on this bug fix.

The issue seems to be that
db_update_cluster() always closes the passed connection in the finally block,
even though the caller may be managing the connection lifecycle.

I will update the code to ensure:

• only internally created connections are closed here
• externally passed connections are left open
• existing tests are not broken and new tests may be added

Please let me know if you want any particular behavior expected here.

[yugalkaushik]

@V4Vikaskumar Thank for your interest but I already made a PR for this issue.