[Enhancement]: Atomicity of data partition decommission operation. #3500
Description
Contact Details
Is there an existing issue for this?
- I have searched all the existing issues
What would you like to be added?
The current decommisison process lacks atomicity. Issues such as datanode or master restarts, network failures, etc., can lead to anomalies in data partition replicas, such as the offline process getting stuck or missing replicas, requiring manual intervention.
Why is this needed?
Ensure data reliability.
Anything else?
No response