[GitHub] [ozone] ChenSammi commented on pull request #2377: HDDS-5296. Replication Manager should not sent replica command to DN which is not in IN_SERVICE opState

[GitBox <gi...@apache.org>]

ChenSammi commented on pull request #2377:
URL: https://github.com/apache/ozone/pull/2377#issuecomment-871849982


   > 
   > 
   > These changes make sense to me, I have pointed out a couple of small things to check inline.
   > 
   > Another scenario for this problem is:
   > 
   >     1. Replication Manager schedules a lot of replication commands on a host where they are queued.
   > 
   >     2. The host is then decommissioned soon after.
   > 
   > 
   > In that case, all those inflight replications are still going to be executed.
   > 
   > Should we also check the DN state in the DN replication command handler, and if the DN is not IN_SERVICE, ignore them?
   > 
   > You can get the DN state inside `ReplicateContainerCommandHander` via:
   > 
   > ```
   > context.getParent().getDatanodeDetails().getPersistedOpState()
   > ```
   > 
   > So if we passed the context.getParent() into `new ReplicationTask(...)` it would be able to make that check when it is executed.
   
   Right. I have added the  DN state check in  ReplicateContainerCommandHandler#handle.


-- 
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.

To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org