sodonnel commented on a change in pull request #262: HDDS-2459 - Change the
ReplicationManager to consider decommission and maintenance states
URL: https://github.com/apache/hadoop-ozone/pull/262#discussion_r351211341
##########
File path:
hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/container/ReplicationManager.java
##########
@@ -97,6 +98,11 @@
*/
private final LockManager<ContainerID> lockManager;
+ /**
+ * Used to lookup the health of a nodes or the nodes operational state.
+ */
+ private final NodeManager nodeManager;
Review comment:
The use of the node manager is not to see the state of the containers, but
to check if a node is alive or not before attempting to use it as the source of
a replication. Eg, the containers can be decommissioned (which is their end
state) but the host is either healthy, stale or dead. We don't want to use a
stale or dead node as a source of the replication hence we need to use the node
manager to get that health state. The proposal by @anuengineer / @nandakumar131
is in place here - the logic expects the DNs to change the container state via
a container report to decommissioned / maintenance for the rest of the logic to
work .
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
