[HDDS-3081] Replication manager should detect and correct containers which don't meet the replication policy

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.0.0
• Fix Version/s: 1.0.0
• Component/s: SCM
• Labels: pull-request-available
• Target Version/s: 1.0.0
• Assignee: Stephen O'Donnell
• Reporter: Stephen O'Donnell
• Created: 26/Feb/20 11:30
• Updated: 03/Jun/20 07:19
• Resolved: 03/Jun/20 07:18

Description

In the current implementation, replication manager does not consider the container placement when checking if a container is healthy. Only the number of replicas are checked.

Now we have network topology available, we should consider whether replication manager should detect and correct mis-replicated containers.

In HDFS, the namenode will not automatically correct mis-replicated containers automatically, except at startup when all blocks are checked.

Issue Links

• links to: GitHub Pull Request #881