zentol commented on a change in pull request #17606:
URL: https://github.com/apache/flink/pull/17606#discussion_r740932666
##########
File path:
flink-tests/src/test/java/org/apache/flink/test/recovery/TaskManagerProcessFailureBatchRecoveryITCase.java
##########
@@ -67,7 +67,7 @@ public void testTaskManagerFailure(Configuration
configuration, final File coord
ExecutionEnvironment env =
ExecutionEnvironment.createRemoteEnvironment("localhost",
1337, configuration);
env.setParallelism(PARALLELISM);
- env.setRestartStrategy(RestartStrategies.fixedDelayRestart(1, 1500L));
+ env.setRestartStrategy(RestartStrategies.fixedDelayRestart(10, 1500L));
Review comment:
The issue depends on when the first heartbeat is sent after the TM went
down (and would then be considered unreachable).
If it happens before the restart, then we properly remove the TM and don't
use it later on.
If it does not happen before the restart, then the job fails later on a
second time while attempting to deploy task to the same TM.
I think this change just ever so slightly adjusts the timings to make it
more common; it can already happen in the current master.
--
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]
