Details
-
Sub-task
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
Description
Currently if there is any exception on the write path, containers will be marked unhealthy, which is a terminal state. Some of these errors do not affect the contents of these containers so they should not actually be marked unhealthy. This jira proposes closing a container normally if an error is encountered, and queueing it for on demand scanning once the close finishes. After this change, the scanner would be the only thing that can mark a container unhealthy.
Attachments
Issue Links
- relates to
-
HDDS-7096 Clarify locations of on-demand scanning
- Open