[jira] [Created] (FLINK-26400) Release Testing: Explicit shutdown signalling from TaskManager to JobManager

["Niklas Semmler (Jira)" <ji...@apache.org>]

Niklas Semmler created FLINK-26400:
--------------------------------------

             Summary: Release Testing: Explicit shutdown signalling from TaskManager to JobManager
                 Key: FLINK-26400
                 URL: https://issues.apache.org/jira/browse/FLINK-26400
             Project: Flink
          Issue Type: Improvement
          Components: Runtime / Coordination
    Affects Versions: 1.15.0
            Reporter: Niklas Semmler
             Fix For: 1.15.0


FLINK-25277 introduces explicit signalling between a TaskManager and the JobManager when the TaskManager shuts down. This reduces the time it takes for a reactive cluster to down-scale & restart.

 

*Setup*

1. Checkout master and install Flink 

{{mvn clean install -Dfast -DskipTests -Dnpm.skip -Dskip.npm}}

2. Go to the build target folder

{{cd build-target }}

3. Add the following line to your flink config to enable reactive mode:

{{taskmanager.host: localhost # a workaround}}
{{scheduler-mode: reactive}}
{{restart-strategy: fixeddelay}}
{{restart-strategy.fixed-delay.attempts: 100}}

4. Create a “usrlib” folder and place the TopSpeedWindowing jar into it

{{mkdir usrlib}}
{{cp ../../../../flink-examples/flink-examples-streaming/target/flink-examples-streaming_2.12-1.15-SNAPSHOT-TopSpeedWindowing.jar usrlib/}}

5. Start the job 

{{bin/standalone-job.sh start  --main-class org.apache.flink.streaming.examples.windowing.TopSpeedWindowing}}

6. Start three task managers

{{bin/taskmanager.sh start}}
{{bin/taskmanager.sh start}}
{{bin/taskmanager.sh start}}

7. Wait for the job to stabilize. The log file should show that three tasks start for every operator.

{{GlobalWindows -> Sink: Print to Std. Out (3/3) (d10339d5755d07f3d9864ed1b2147af2) switched from INITIALIZING to RUNNING.}}

 

*Test*

Stop one taskmanager

{{bin/taskmanager.sh stop}}

Success condition: You should see that the job cancels and re-runs after a few seconds. In the logs you should find a line with the text “The TaskExecutor is shutting down”.

 

*Teardown*

Stop all taskmanagers and the jobmanager:

{{bin/standalone-job.sh stop}}
{{bin/taskmanager.sh stop-all}}



--
This message was sent by Atlassian Jira
(v8.20.1#820001)