Stephen O'Donnell created HDDS-7998:
---------------------------------------
Summary: Synchronize on containerInfo in ReplicationManager and
MoveManager
Key: HDDS-7998
URL: https://issues.apache.org/jira/browse/HDDS-7998
Project: Apache Ozone
Issue Type: Sub-task
Components: SCM
Reporter: Stephen O'Donnell
Assignee: Stephen O'Donnell
Keeping in line with the Legacy Replication Manager, we should synchronize on
the containerInfo currently being processed within RM and the new MoveManager.
ICR and FCR processing already synchronizes on containerInfo as it processes
the container, and in order to get a consistent view of the container as it is
processed, Legacy ReplicationManager also synchronized on containerInfo.
By ensuring ICR / FCR, RM and MoveManager synchronize on the containerInfo
object, we can ensure that each processing area gets a consistent view of the
container pendingOp, replicas and state as it processes the container, and
avoid any unexpected consequences around concurrently modifying the container
state in SCM.
Replication manager is now split into 2 stages. The "check" stage, and the
unhealthy (under, over, mis-replicated) processing stage. Therefore we need to
synchronized in two places inside RM.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
