[ https://issues.apache.org/jira/browse/FLINK-19570?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Zhu Zhu closed FLINK-19570. --------------------------- Resolution: Fixed Fixed via f2d70c85b659074342db93bd903b2c9e91cfb749 0c373822e31562c6f0d65505c0b95ea1b166c093 > Execution graph related tests are possibly broken due to registering > duplicated ExecutionAttemptID > -------------------------------------------------------------------------------------------------- > > Key: FLINK-19570 > URL: https://issues.apache.org/jira/browse/FLINK-19570 > Project: Flink > Issue Type: Bug > Components: Runtime / Coordination, Tests > Affects Versions: 1.12.0 > Reporter: Zhu Zhu > Assignee: Zhu Zhu > Priority: Critical > Labels: pull-request-available > Fix For: 1.12.0 > > > Since FLINK-17295, many tests encounters unexpected global failure due to > registering duplicated ExecutionAttemptID. Although these tests do not appear > to be broken yet, they are potentially broken/unstable. And it further blocks > to rework these tests to be based on the new scheduer (FLINK-17760). > Below is a sample error which happens in > ExecutionTest#testAllPreferredLocationCalculation(): > {code:java} > 2194 [main] INFO org.apache.flink.runtime.executiongraph.ExecutionGraph [] - > Job (unnamed job) (a22afb832b5f94b075d7ffb32fbc9023) switched from state > CREATED to FAILING. > java.lang.Exception: Trying to register execution Attempt #0 (TestVertex > (1/1)) @ (unassigned) - [CREATED] for already used ID > a22afb832b5f94b075d7ffb32fbc9023_146968a4de2df0b2fef1e4b2e8297993_0_0 > at > org.apache.flink.runtime.executiongraph.ExecutionGraph.registerExecution(ExecutionGraph.java:1621) > [classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionVertex.<init>(ExecutionVertex.java:181) > [classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionJobVertex.<init>(ExecutionJobVertex.java:211) > [classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionJobVertex.<init>(ExecutionJobVertex.java:139) > [classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionGraphTestUtils.getExecutionJobVertex(ExecutionGraphTestUtils.java:448) > [test-classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionGraphTestUtils.getExecutionJobVertex(ExecutionGraphTestUtils.java:419) > [test-classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionGraphTestUtils.getExecutionJobVertex(ExecutionGraphTestUtils.java:411) > [test-classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionGraphTestUtils.getExecutionJobVertex(ExecutionGraphTestUtils.java:452) > [test-classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionGraphTestUtils.getExecution(ExecutionGraphTestUtils.java:477) > [test-classes/:?] > at > org.apache.flink.runtime.executiongraph.ExecutionTest.testAllPreferredLocationCalculation(ExecutionTest.java:298) > [test-classes/:?] > at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) > ~[?:1.8.0_261] > at > sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62) > ~[?:1.8.0_261] > at > sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) > ~[?:1.8.0_261] > at java.lang.reflect.Method.invoke(Method.java:498) ~[?:1.8.0_261] > at > org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:50) > [junit-4.12.jar:4.12] > at > org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12) > [junit-4.12.jar:4.12] > at > org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:47) > [junit-4.12.jar:4.12] > at > org.junit.internal.runners.statements.InvokeMethod.evaluate(InvokeMethod.java:17) > [junit-4.12.jar:4.12] > at org.junit.rules.TestWatcher$1.evaluate(TestWatcher.java:55) > [junit-4.12.jar:4.12] > at org.junit.rules.RunRules.evaluate(RunRules.java:20) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner.runLeaf(ParentRunner.java:325) > [junit-4.12.jar:4.12] > at > org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:78) > [junit-4.12.jar:4.12] > at > org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:57) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner$3.run(ParentRunner.java:290) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner$1.schedule(ParentRunner.java:71) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner.runChildren(ParentRunner.java:288) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner.access$000(ParentRunner.java:58) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner$2.evaluate(ParentRunner.java:268) > [junit-4.12.jar:4.12] > at > org.junit.rules.ExternalResource$1.evaluate(ExternalResource.java:48) > [junit-4.12.jar:4.12] > at org.junit.rules.RunRules.evaluate(RunRules.java:20) > [junit-4.12.jar:4.12] > at org.junit.runners.ParentRunner.run(ParentRunner.java:363) > [junit-4.12.jar:4.12] > at org.junit.runner.JUnitCore.run(JUnitCore.java:137) > [junit-4.12.jar:4.12] > at > com.intellij.junit4.JUnit4IdeaTestRunner.startRunnerWithArgs(JUnit4IdeaTestRunner.java:69) > [junit-rt.jar:?] > at > com.intellij.rt.junit.IdeaTestRunner$Repeater.startRunnerWithArgs(IdeaTestRunner.java:33) > [junit-rt.jar:?] > at > com.intellij.rt.junit.JUnitStarter.prepareStreamsAndStart(JUnitStarter.java:220) > [junit-rt.jar:?] > at com.intellij.rt.junit.JUnitStarter.main(JUnitStarter.java:53) > [junit-rt.jar:?] > {code} > This is because these tests improperly creates > Execution/ExecutionVertex/ExecutionJobVertex from an existing ExecutionGraph > which already contains the Execution/ExecutionVertex/ExecutionJobVertex. > FLINK-17295 reveals this problem because it makes the attemptID no longer > random. > Below is an example of improper ExecutionJobVertex creation in > ExecutionGraphTestUtils#getExecutionJobVertex(): > {code:java} > JobGraph jobGraph = new JobGraph(ajv); > jobGraph.setScheduleMode(scheduleMode); > ExecutionGraph graph = TestingExecutionGraphBuilder > .newBuilder() > .setJobGraph(jobGraph) > .setIoExecutor(executor) > .setFutureExecutor(executor) > .build(); > > graph.start(ComponentMainThreadExecutorServiceAdapter.forMainThread()); > return new ExecutionJobVertex(graph, ajv, 1, > AkkaUtils.getDefaultTimeout()); > {code} > We should get rid of such improper usages. > Therefore, I would like to change these tests to get existing > Execution/ExecutionVertex/ExecutionJobVertex from the generated > ExecutionGraph, instead of invoking their constructors to create new ones. -- This message was sent by Atlassian Jira (v8.3.4#803005)