恢复计划器持续报告"文件未找到",并在对象被删除后无法自动恢复。
作者: aubdiy创建于 2026年9月29日更新于 2026年9月30日
标签S-confirming
- Problem description (Description) The Heal scheduler in the RustFS cluster continuously prints the task_failed error, indicating Storage error: File not found. The key exception is that even after manually removing the corresponding .incomplete temporary object through the rc object remove, the Heal scheduler continues to retry the task and reports an error, failing to correctly detect that the object has been deleted and clean up the task queue, resulting in a log screen. At the same time, a large number of http_request_inflight_slow warnings appear, and the ReadVersion RPC call takes more than 5 seconds, which is suspected to be related to the Heal task being blocked or the metadata being inconsistent.
- Environment information (Environment) Components: rustfs-server / heal-scheduler Deployment method: Kubernetes (bd-rustfs-4 namespace) Affected Bucket/Object: dl-ods/adt_callback/.../.incomplete/... Time range: 2026-09-24T09:49:13Z, continuously occurring
内容来源: rustfs/rustfs