PatrickRen commented on pull request #16951:
URL: https://github.com/apache/flink/pull/16951#issuecomment-904710561
Thanks @zentol for the review!
After a second thought I'm kinda doubting whether we need this check in
```triggerTaskManagerFailover```. Initially I was thinking we need to make sure
that the job indeed suffers a failover after terminating the TM, so I used
```waitForJobStatus``` to wait until job getting into FAILED, FAILING or
RESTART status after killing the TM. Actually this is not necessary since we
always terminate the only TM and start a new one to trigger failover, so the
job will definitely suffer a failover.
I'd like to just remove the ```waitForJobStatus``` in
```MiniCluster#triggerTaskManagerFailover``` instead of using metric to detect
restart:
```java
@Override
public void triggerTaskManagerFailover(JobClient jobClient, Runnable
afterFailAction) throws Exception {
terminateTaskManager();
afterFailAction.run();
startTaskManager();
}
```
What do you think @zentol @AHeise ?
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
