DanielLeens commented on issue #13323: URL: https://github.com/apache/seatunnel/issues/13323#issuecomment-6983651480
Confirmed the retry-budget reset at the **SubPlan reconstruction boundary**, using actual constructors and `canRestorePipeline()` on `89f91a89af15f33d97a942d8b65f4fc1b0c80bf5` (Java 11.0.27). Steps: 1. Construct a real SubPlan with `job.retry.times=3` and `JobMaster.isNeedRestore()=true`. 2. Set its private restore counter to 3 to represent an exhausted prior generation, and retain a restore cause. `canRestorePipeline()` is false. 3. Construct a second SubPlan with the **same job configuration and same IMap state**, without carrying over process-local fields. 4. The new SubPlan has counter 0, empty restore causes, and `canRestorePipeline()` returns true. Observed: ```text old count=3/canRestore=false reconstructed count=0/canRestore=true, causes empty ``` [Executable Java probe and exact build/run commands](https://github.com/apache/seatunnel/issues/13314#issuecomment-6983602245): see `retryBudget()` in `StateRepros.java`. This deliberately seeds the exhausted counter rather than running three real job failures. It proves the reconstruction behavior only; it is **not** a multi-node master-failover experiment or evidence of an observed infinite crash loop. -- 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]
