fix(framework): Make message delivery safe for multiple LinkState replicas

[psfoley]

Issue

Description

Related issues/PRs

Proposal

Explanation

Checklist

• Implement proposed change
• Write tests
• Update documentation
• Make CI checks pass
• Ping maintainers on Slack (channel #contributions)

Any other comments?

[Copilot]

Pull request overview

This PR updates the SQL-backed LinkState message retrieval logic to prevent duplicate delivery when multiple SuperLink/LinkState replicas concurrently pull the same instruction or reply from a shared SQLite database.

Changes:

• Make get_message_ins claim instructions more safely by only updating rows that are still undelivered (delivered_at = '') and relying on RETURNING to fetch the claimed rows.
• Make get_message_res atomically claim eligible reply messages via a single UPDATE ... RETURNING instead of SELECT then UPDATE.
• Add concurrency tests to validate that concurrent replicas cannot both claim the same instruction/reply.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 4 comments.

File	Description
framework/py/flwr/server/superlink/linkstate/sql_linkstate.py	Tightens delivery semantics by atomically claiming replies and preventing double-claim of instructions via delivered_at guards.
framework/py/flwr/server/superlink/linkstate/linkstate_test.py	Adds multi-replica concurrency tests for unique claiming of instruction and reply messages.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

This PR makes SQL-backed message delivery safe under contention when multiple SuperLink LinkState replicas share the same SQLite DB, by turning “read then mark delivered” into an atomic “claim and return” operation.

Changes:

• Atomically “claim” instruction messages in get_message_ins via UPDATE ... SET delivered_at ... RETURNING (including a deterministic subquery when limit is set).
• Atomically “claim” reply messages in get_message_res via UPDATE ... SET delivered_at ... RETURNING.
• Add concurrency tests validating uniqueness of claims across two SqlLinkState instances sharing the same database file.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
framework/py/flwr/server/superlink/linkstate/sql_linkstate.py	Switches message fetching to atomic claim-and-return updates to prevent duplicate delivery across replicas.
framework/py/flwr/server/superlink/linkstate/linkstate_test.py	Adds multi-threaded tests exercising concurrent message claiming across two SQL linkstate replicas.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[Copilot]

Pull request overview

This PR updates the SQLite-backed SqlLinkState message retrieval logic to make instruction/reply delivery safe under concurrent access from multiple LinkState replicas, and adds regression tests to validate unique claiming behavior under contention.

Changes:

• Make get_message_ins atomically “claim” messages by UPDATE ... SET delivered_at ... RETURNING * (with a follow-up re-read after validity checks).
• Make get_message_res atomically “claim” reply messages via UPDATE ... RETURNING *, removing the non-atomic select-then-update flow.
• Add multi-threaded tests using a shared SQLite file to ensure unique claiming and work distribution across two replicas.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
framework/py/flwr/server/superlink/linkstate/sql_linkstate.py	Switches message retrieval to atomic claim semantics using UPDATE ... RETURNING for safe multi-replica delivery.
framework/py/flwr/server/superlink/linkstate/linkstate_test.py	Adds concurrent replica tests to validate unique claiming and contention behavior against a shared SQLite DB.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

You can also share your feedback on Copilot code review. Take the survey.

[msheller]

@danieljanes @psfoley , this implementation looks correct to me. I'm in favor of merging it. Once we have this and the running state issue addressed, we should run another analysis on concurrency risks.