Repository navigation
Reuse target metadata plan to speed up live delta comparison - #70
Open
gspark-etri wants to merge 8 commits into
Open
gspark-etri wants to merge 8 commits into
gspark-etri wants to merge 8 commits into
Conversation
This was referenced Oct 30, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Every live resident update invokes onggi diff, which currently rebuilds the target's metadata plan twice. Reuse one validated target plan for both sorted extents and distinct new stored-byte accounting, preserving hash checks, stale-locator/index fallback and JSON/API behavior.
Measured on the same immutable8B POSIX-store fixtures with Go1.27.3/GOMAXPROCS4: four ref directions, six alternating pairs each. All48 CLI outputs are byte-identical within their ref pair. Median CLI time fell8.30–13.76% (about24–40ms). This isolates CPU comparison; NPU/cold-load/end-to-end latency and S3 performance improvements are unclaimed.
Regression RED observed two target-locator reads; GREEN expects one. Legacy/identical/empty artifacts, damaged target locator and GC-repack Diff fallback are covered. Linux full go test -race ./..., go vet ./... and gofmt pass. Fresh read-only review found one test import-order issue, fixed and gates rerun; no production/harness correctness issue found.
Protocol and scope, all samples/source and binary hashes/ref digests. Public bench/diff.py reproduces the interleaved comparison. No NPU allocated or shared SDK/driver changed. Base is main; independent of the SDK migration/doc PRs.
Final head358c75529808beecf854d7b65856e519accd3cad CI passed:gofmt,vet,fullrace suite,LinuxCLI build, and real S3 conformance/byte-identical round trips against SeaweedFS and VersityGW. Source/harness bytes are unchanged from the locally validated93fa367. The remote source worktree is clean; owned test/build/benchmark processes are absent. No NPU allocation was created. AI-assisted implementation and verification with Codex.