[ENH] When the log offset is behind sysdb, this can repair it safely.

[rescrv]

Description of changes

This adds a tool to make the log offset match sysdb. Then a test to make sure it cannot move things
backwards breaking invariants.

Test plan

Added a test.

• Tests pass locally with pytest for python, yarn test for js, cargo test for rust

Documentation Changes

N/A

[github-actions]

Reviewer Checklist

Please leverage this checklist to ensure your code review is thorough before approving

Testing, Bugs, Errors, Logs, Documentation

• Can you think of any use case in which the code does not behave as intended? Have they been tested?
• Can you think of any inputs or external events that could break the code? Is user input validated and safe? Have they been tested?
• If appropriate, are there adequate property based tests?
• If appropriate, are there adequate unit tests?
• Should any logging, debugging, tracing information be added or removed?
• Are error messages user-friendly?
• Have all documentation changes needed been made?
• Have all non-obvious changes been commented?

System Compatibility

• Are there any potential impacts on other parts of the system or backward compatibility?
• Does this change intersect with any items on our roadmap, and if so, is there a plan for fitting them together?

Quality

• Is this code of a unexpectedly high quality (Readability, Modularity, Intuitiveness)

[propel-code-bot]

Enable Safe Repair of Log Offset to System Database State

This PR adds tooling and backend changes to safely synchronize a collection's log offset with the system database (sysdb) if it has fallen behind, preventing illegal backward movement of the offset. The changes span proto, Rust, and Go domains, including a new Rust tool for repair, a strong server-side test for log invariants, and sysdb interface updates to surface required information.

Key Changes:
• Introduced a standalone Rust tool (bin/chroma-update-collection-log-offset.rs) to synchronize log offset from sysdb for a collection.
• Added a comprehensive Rust test (update_collection_log_offset_never_moves_backwards) ensuring log offsets cannot be moved backward, maintaining WAL invariants.
• Updated sysdb proto (CheckCollectionsResponse) and Go backend to provide log positions alongside deleted status for collections.
• Refactored Go coordinator, service, and catalog layers to wire log_position into the sysdb API and through to check responses.

Affected Areas:
• Rust log-service (tests, CLI tool)
• Go sysdb coordinator, service, and proto
• IDL (Proto file, coordinator.proto)

Potential Impact:

Functionality: Users and operators can now repair a collection's log offset in case of desynchronization, with safeguards to prevent violating offset monotonicity.

Performance: Negligible impact; the repair tool and proto expansion add minimal runtime complexity.

Security: No new externally-exposed vectors; repair tool assumes trusted access and sysdb/reflection pairing.

Scalability: Neutral-changes make operations safer, even at large scale, but fundamentally do not alter scaling characteristics.

Review Focus:
• Correctness of monotonicity enforcement during log offset updates in Rust service and associated test coverage.
• Go sysdb and proto contract-ensure all usages of CheckCollections and new log_position semantics are consistently applied.
• Backward compatibility with systems/components that previously expected CheckCollectionsResponse to include only deleted status.
• CLI usability and error reporting in chroma-update-collection-log-offset tool.

Testing Needed

• Manual invocation of the new chroma-update-collection-log-offset tool against intentionally skewed log/sysdb states.
• Run the supplied Rust test suite to check log offset update invariants.
• Verify interoperability after proto/grpc changes between Go sysdb and Rust log service.

Code Quality Assessment

rust/log-service/src/lib.rs: Added a well-structured async test ensuring strict invariants. Change is sound and idiomatic.

rust/log-service/src/bin/chroma-update-collection-log-offset.rs: Tool is concise, idiomatic, and includes clear error-handling.

go/pkg/sysdb/coordinator/table_catalog.go: Refactor maintains style and handles new log_position output safely.

go/pkg/sysdb/grpc/collection_service.go: Proto method changes are correctly propagated; code is still readable.

go/pkg/sysdb/coordinator/coordinator.go: Simple function signature update; transparent change.

idl/chromadb/proto/coordinator.proto: Proto update is minimal, clear, backward compatible as long as clients properly ignore unknown fields.

Best Practices

Testing:
• Regression and invariant unit tests provided for critical mutation path.

Backward Compatibility:
• Proto changes are additive; but recommend confirming client behaviors.

Validation:
• Input argument and state validation in CLI tool
• Comprehensive assertion-based testing for core invariants

Potential Issues

• External consumers of CheckCollectionsResponse may break if they don't ignore the new log_position field; confirm downstream compatibility.
• CLI repair tool requires precise host arguments and is not self-discovering-misuse could go quietly unnoticed without clear logs.
• If sysdb and log service are at incompatible schema or versions during rolling upgrades, log offset repair may fail or behave unexpectedly.

This summary was automatically generated by @propel-code-bot

[Sicheng-Pan]

nit: "deleted collection"?