SEZ9 commented on PR #12173: URL: https://github.com/apache/seatunnel/pull/12173#issuecomment-5707365832
@Rangsh thanks for the status on `7948e2640` — sorry this sat unanswered. Going through your points: - **F4 (docs)**: agreed, confirmed. With the notes added in `909f878` and removed in `9a6077c58`, the head no longer touches `docs/en|zh/engines/zeta/benchmark.md`, so there is no doc mismatch left. Closed. - **F1/F3 (`IMapJobGrowthBenchmarkWorkload`)**: understood, and thanks for correcting my "once-per-fork" wording. The sampling points you list (`initialStoredJobCount=0`: first iteration + trial tear-down; `initialStoredJobCount=1000`: trial tear-down only, everything else resident-only) are an acceptable trade-off given that a mid-trial full-WAL reload at pressure=1000 OOMs the JMH fork. I'm fine keeping the single-representative-key sample here and treating full-batch widening as follow-up work in #12266. Closed for this PR. - **F2/F5 (`storeFinishedPipelineMetrics`)**: that resolves my concern. Wrapping `get → merge → put` in the cluster-wide `finishedJobMetricsImap.lock(jobId)` / `unlock(jobId)` with `try`/`finally` (since `ca82f95`) is exactly what I was after, and the contract-only Javadoc (lock + merge + single TTL `put`, no "in memory" / `computeIfAbsent` wording) reads correctly now. Closed. - **F6 (`JobHistoryServiceFinishedMetricsTest`)**: your last line got truncated in the comment, but `storeFinishedPipelineMetricsRejectsNullMetrics` for the null path plus the new-job/merge/distinct-tag tests asserting the TTL-preserving `put(..., 1440, MINUTES)` and explicit `lock`/`unlock` cover what I asked for. Closed. So from my side F1–F6 are all resolved on `7948e2640` and I have no further code asks; no new push is needed. The only thing still standing is the earlier changes-requested review that predates the `9a6077c58` doc fix — I'll follow up on getting that re-reviewed so the PR is unblocked. <!-- streview-comment:1101 --> -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
