fix(palace): postgres backend is a multi-process-writer backend

[jphein]

Summary

Post-sync production regression, caught in the wrap-up sweep: upstream v3.7's process-lifetime MCP writer lease (_acquire_mcp_writer_lock) exempts only {pgvector, qdrant} via _MULTI_PROCESS_WRITER_BACKENDS. The fork's postgres backend wasn't in the list, so the daemon's embedded MCP server took the lease at startup — and every mempalace mine subprocess the daemon spawns for POST /mine self-refused with palace … is held by PID <daemon> (main.py). 45 refused transcript ingests in hook.log since the 3.7.0 deploy.

postgres coordinates concurrent clients server-side — exactly the documented opt-out condition (same as pgvector). One-word fix + regression assertion in the existing ownership test.

Test plan

• tests/test_palace.py extended: backend_requires_single_writer('postgres') is False
• test_palace.py + test_mcp_server.py suites green locally
• Post-merge: deploy to familiar, verify a real POST /mine returns returncode 0 and hook.log refusals stop

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Improved PostgreSQL backend behavior for multi-process write access.
  • PostgreSQL no longer requires exclusive single-writer ownership for the lifetime of a process.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: f9203c54-4aa6-456e-9f9d-756eb9f3f5e8

📥 Commits

Reviewing files that changed from the base of the PR and between b033c10 and 6b75db6.

📒 Files selected for processing (2)

• mempalace/palace.py
• tests/test_palace.py

---

📝 Walkthrough

Walkthrough

The change classifies PostgreSQL as a multi-process writer backend. Tests verify that PostgreSQL does not require single-writer ownership.

Changes

PostgreSQL writer coordination

Layer / File(s)	Summary
Backend classification and validation mempalace/palace.py, tests/test_palace.py	The backend list includes postgres. The test verifies that backend_requires_single_writer("postgres") returns False.

Estimated code review effort: 2 (Simple) | ~10 minutes

Suggested reviewers: igorls, bensig

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely identifies the PostgreSQL backend concurrency fix, which is the main change in the pull request.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/postgres-multiprocess-writer

---

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.}