Toroidals commented on issue #13114: URL: https://github.com/apache/hudi/issues/13114#issuecomment-2791372609
> > 20250408130311013_20250408132310766_0.parquet > > yes, this is a archived timeline layer 0 file, the two timestamps in the fiile name are start/end instant times archived from the active time. Maybe the write to archived timeline succeed but the deletion of active timeline files failed, I will log in issue to address it, would resolve it in 1.0.2 release: https://issues.apache.org/jira/browse/HUDI-9286 When using version 0.15.0, the job executed normally, and rollback worked correctly when the job failed. After upgrading to 1.0.0, without making significant changes to the write logic, the job still runs normally, but in most failure cases, rollback no longer works. I ran around 30 jobs, and more than half initially failed with the following error: Caused by: org.apache.hadoop.ipc.RemoteException: File does not exist: /apps/hive/warehouse/hudi.db/hudi_ert_cmf_inf_po_locations_cdc/.hoodie/metadata/files/.files-0000-0_20250410020542786.log.1_0-1-0 (inode 9447634622) Holder DFSClient_NONMAPREDUCE_-17041629_40 does not have any open files. This led to subsequent errors and eventually caused rollback failure. **Stacktrace** https://gist.github.com/Toroidals/49742f9605599d3ec8bec60b76127cec -- 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]
