aglinxinyuan commented on PR #6797: URL: https://github.com/apache/texera/pull/6797#issuecomment-5076533156
Pushed b67bf3acd7 to fix a noise regression the WARN backstop itself introduced — the CI logs were full of this on every `ActorSystem` creation: ``` [ERROR] [EventStream(pekko://CheckpointSubsystemSpec-test)] unknown pekko.loglevel WARN org.apache.pekko.event.Logging$LoggerException ``` **Root cause:** `cluster.conf` forwards `TEXERA_SERVICE_LOG_LEVEL` into `pekko.loglevel`, and pekko shares loguru's vocabulary (`WARNING`), not logback's (`WARN`). On an unknown name pekko prints a `LoggerException` stack trace per `ActorSystem` and falls back to `ERROR` — one per spec in both amber jobs. **Why not just set `WARNING` in the env:** the same variable drives the logback root level in `logback.xml`, and logback silently falls back to **DEBUG** on names it doesn't know — that would flood instead of quiet. The two vocabularies are irreconcilable at the env level. **Fix:** normalize at the single choke point. `PekkoConfig.pekkoConfig` (which every `ActorSystem`, prod and test, receives via `AmberRuntime.pekkoConfig`) now rewrites `pekko.loglevel` through `normalizePekkoLogLevel` (`WARN → WARNING`, `TRACE`/`ALL` → `DEBUG`), so the one env knob drives both systems in either spelling. Test-first: added an invariant test to `PekkoConfigSpec` — the resolved `pekko.loglevel` must be in pekko's accepted set whatever the env holds — verified red under `TEXERA_SERVICE_LOG_LEVEL=WARN` before the fix, green after (7/7 with WARN, 65/65 Config module with env unset), plus unit tests for the mapping itself. -- 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]
