[GitHub] [flink-kubernetes-operator] wangyang0918 edited a comment on pull request #109: [FLINK-26577] Always trigger a savepoint when upgrade mode changed to last-state

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

wangyang0918 edited a comment on pull request #109:
URL: https://github.com/apache/flink-kubernetes-operator/pull/109#issuecomment-1078591180


   > Maybe the only thing I would improve now that I am playing around with it is that if we have HA enabled for a job and the job is already running , we probably should not trigger a savepoint when switching to last-state.
   > 
   > Do you see any corner cases with this logic?
   
   Actually, I also have this improvement in my mind at the very begging and not sure about whether to introduce this. Because we could not guarantee that there is a successful checkpoint in the ConfigMap. You could argue that the state will also be lost when JobManager failed over. But changing the upgrade mode is not a frequent operation, I think triggering a savepoint is reasonable. 
   
   Another concern is we could allow cross major version upgrading(1.14 -> 1.15), but the HA ConfigMap might be compatible. 
   
   WDYT?


-- 
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: commits-unsubscribe@flink.apache.org

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