[HDDS-6518] EC: Add rpc for EC recovery in replication service

• Type: Sub-task
• Status: Resolved
• Priority: Major
• Resolution: Won't Do
• Labels: pull-request-available
• Assignee: Kaijie Chen
• Reporter: Kaijie Chen
• Created: 28/Mar/22 03:39
• Updated: 11/May/22 08:17
• Resolved: 11/May/22 08:17

Description

When replicating EC containers, we need more rpc than copy container.

CoordinatorDN -> HealthyDN

This patch adds support for ContainerCommand in replication service.

Server Side:

1. Added command handler to dispatch container command requests.

Client Side:

1. Added readContainer() to get number of total blocks to list.
2. Added listBlock() to list blocks in a container.
3. Added readChunk() to read the data from healthy datanodes.

CoordinatorDN -> TargetDN

The CoordinatorDN should push the chunks to the TargetDN after computing missing blocks.
WriteChunkProto without data means the end of the container.

Push is chosen for following reasons:

1. CoordinatorDN does not need to hold the missing blocks for long.
2. CoordinatorDN can controll the progress of the recovery.

Server Side:

1. Added dummy command handler for push chunk and start recovery.

Client Side:

1. Added writeChunk() to push chunks to target datanodes.

Issue Links

• links to: GitHub Pull Request #3351