> the idea of the patch was to keep the supervisor in a state similar to what 
> it would have been in if it were stuck in the current lifecycle start, while 
> allowing the rest of the lifecycle to start

Got it. Makes sense. 

The reason I commented about `InitNotice` was, I'm not sure 
`initializationExec` is needed. Maybe we can use `scheduledExec` instead and 
cancel the scheduledFuture once `tryInit` succeeds? Or we can submit a runnable 
task to `exec` which calls `RetryUtils.retry()`.

[ Full content available at: 
https://github.com/apache/incubator-druid/pull/6383 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to