Repository navigation
test: benchmark fivefold ten-commit histories directly on Aliyun ECS - #205
Merged
Merged
Conversation
…paces Add a bounded history-large workload with ten real commits prepared before measurement: one cold population and nine complete module rewrites, each changing 16,384 distinct 8 KiB files (128 MiB) with deterministic pseudorandom bytes. Preserve the existing 100k-file four-scenario workload. The history profile uses 17,026 logical files to stay within the frozen server cumulative source ceilings. Replay raw Git commit hashes, parents, trees, manifest differences and matching Git/ScorpioFS publications. Require all ten retained views, descriptor and dirty-upper checks, 60 fair A/B samples, 10 independent diagnostic samples and 525 full oracles without extending the four-hour cloud window. Add negative cases for missing or repeated commits, v10 ordering and altered receipts. Production Rust, pinned server/client sources and cloud resource topology remain unchanged. Plans, generated fixtures and evidence stay local.
… ECS Replace the Aliyun Actions/SSH installer with direct Cloud Assistant execution on a Terraform-owned Ubuntu 24.04 host. Deliver exact pinned Git objects and collect only validated evidence through private OSS using IMDSv2 instance credentials; retain the original four-hour window, terminate writers before object deletion, destroy the stack and audit compute, disks, network, OSS, RAM and local state. Support smoke, medium and history-large tiers without a GitHub runner. Grow history-large to 85,122 logical files and ten real commits. Rotate five cohorts of 33 modules: each incremental commit rewrites 16,896 distinct files / 132 MiB while preserving unchanged subtrees and all retained views. Keep every production quota unchanged; the frozen server exact-tree reuse contract bounds source references to 246,425 / 262,144. Require 213,559,413,282 bytes of free work disk, real source/build/instance provenance and complete native correctness oracles. Production Rust and the frozen server/client input revisions are unchanged. Plans, source packages, generated fixtures and evidence remain outside the repository.
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
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
1 participant
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.
Run the Aliyun benchmark directly on Terraform-owned ECS through Cloud Assistant. Remove the Actions runner and SSH installer, deliver exact pinned source objects through private OSS using an IMDSv2 instance role, and replay the allowlisted evidence locally. Provide independent smoke, medium and history-large tiers; retain one immutable four-hour window, stop writers before removing objects, and destroy/audit compute, disks, network, OSS, RAM and Terraform state.
The history workload grows to 85,122 files and ten real commits prepared before measurement. Each incremental commit rotates a 33-module cohort and changes 16,896 distinct files (132 MiB). Verify unchanged files, directory aliases, all ten retained views, descriptors and dirty upper layers, and replay commit parents/trees/raw object hashes. Require 60 fair A/B samples, 10 independent diagnostic samples and 525 complete oracles. Exact unchanged Git tree reuse bounds retained source references to 246,425 / 262,144 with every production limit unchanged. The 300 GiB ESSD work disk must admit a 213,559,413,282-byte reservation.
Validation: 174 bounded Python checks (17 platform skips), 32 direct-controller/source-transport checks, 10 Terraform mock contracts, changed Python syntax, whitespace, production-unchanged checks, and independent exact-source review passed. Source transport tests preserve real Git object IDs and clean shallow checkouts. Native cloud execution follows this pinned harness; no new performance measurement is claimed here. Production Rust and the frozen server/client input revisions are unchanged. Plans, generated fixtures, state, logs and evidence stay outside Git. Both commits have valid GitHub signatures.