LuciferYang opened a new issue, #10237: URL: https://github.com/apache/paimon/issues/10237
### Search before asking - [X] I searched in the [issues](https://github.com/apache/paimon/issues) and found nothing similar. ### Paimon version master ### Compute Engine JavaAPI / Flink / Spark (DDL) ### Minimal reproduce step `FileSystemBranchManager.dropBranch` catches the `IOException` thrown by `fileIO.delete(branchPath, true)`, logs it at INFO level, and returns normally. Because the method returns void and throws nothing, the drop-branch DDL reports success even though the branch directory and all of its data are still on the filesystem. Run `ALTER TABLE t DROP BRANCH b` (filesystem catalog) where the branch directory delete fails, for example due to a permission or filesystem error. ### What doesn't meet your expectations? The DDL succeeds, so the user believes the branch is gone, but its directory and data remain and storage is not reclaimed. A later `CREATE BRANCH b` with the same name then fails with "Branch name 'b' already exists", which is hard to explain to someone who was told the drop succeeded. The INFO message even says "Please try again", yet no error reaches the caller to prompt a retry. When the branch directory cannot be deleted, the DDL should fail so the caller knows the branch was not dropped. This also matches `createBranch`, `renameBranch`, and `fastForward` in the same class, and the catalog-backed `CatalogBranchManager.dropBranch`, which all propagate the failure. ### Anything else? _No response_ ### Are you willing to submit a PR? - [X] I'm willing to submit a PR! -- 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]
