Rangsh commented on PR #12173:
URL: https://github.com/apache/seatunnel/pull/12173#issuecomment-5692790519
@SEZ9 status on `7948e2640` (no new push):
- **F4 (docs)**: confirmed — the growth-scenario notes remain removed on
`7948e2640` (removed in `9a6077c58`); no further doc change.
- **F1/F3 (`IMapJobGrowthBenchmarkWorkload`)**: unchanged by design; **not**
strictly once-per-fork.
Durable `evict`+`loadAll` reloads one representative (last) key only (not
full-batch / not per-iteration). Sampling points:
- `initialStoredJobCount=0`: first iteration **and** trial tear-down
- `initialStoredJobCount=1000`: trial tear-down only
Every other iteration is resident-only. Mid-trial full-WAL reload under
pressure=1000 OOMs the Diagnostics/JMH fork; full-batch widening tracked in
#12266.
- **F2/F5 (`storeFinishedPipelineMetrics`)**: Javadoc has been tightened to
contract-only wording (lock + merge + single TTL `put`; no “in memory” /
historical `computeIfAbsent` wording). On locking: it does **not** rely on
per-JobMaster `metricsLock`; `get → merge → put` is wrapped in cluster-wide
`finishedJobMetricsImap.lock(jobId)` / `unlock(jobId)` (`try`/`finally`) since
`ca82f95`.
- **F6 (`JobHistoryServiceFinishedMetricsTest`)**: covered, not deferred.
Null path: `storeFinishedPipelineMetricsRejectsNullMetrics`.
TTL-preserving path: new-job / merge / distinct-tag tests all verify
`put(..., FINISHED_JOB_EXPIRE_MINUTES, MINUTES)` (1440), plus `lock`/`unlock`.
Happy for you to re-check against this head and close out.
--
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]
