xintongsong commented on issue #11323: [FLINK-16439][k8s] Make 
KubernetesResourceManager starts workers using WorkerResourceSpec requested by 
SlotManager
URL: https://github.com/apache/flink/pull/11323#issuecomment-601058005
 
 
   Thanks for the review, @wangyang0918.
   
   Regarding your concern, I think the current behavior does align with the 
original one. If a TM is terminated exceptionally, whether we request a new pod 
is not decided by the terminated TM, but decided by whether SlotManager has 
more pending slots (pending workers after the changes) than that of 
ResourceManager's pending TMs/pods. In case of JM failover, if the new SM has 
not requested any slots/workers from RM, then there will be no pending 
slots/workers, and RM will not request new TM/pod anyway. If the new SM has 
requested new worker, then there should be corresponding `WorkerResourceSpec` 
in `podWorkerResources`.

----------------------------------------------------------------
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

Reply via email to