[IGNITE-18549] Fix timeouts of CdcKafkaReplicationTest#afterTest

• Type: Test
• Status: Open
• Priority: Minor
• Resolution: Unresolved
• Component/s: extensions
• Labels: IEP-59 ise
• Assignee: Unassigned
• Reporter: Ilya Shishkov
• Created: 13/Jan/23 15:02
• Updated: 16/Jan/23 07:18

Description

Currently, there are flaky timeouts in an #afterTest method of CdcKafkaReplicationTest, CdcKafkaReplicationAppsTest.

Sometimes, tests stuck on waiting of topics deletion and thread dump is printed into log:

[2023-01-13T08:13:22,320][INFO ][after-test-timeout-org.apache.ignite.cdc.kafka.CdcKafkaReplicationTest][] org.apache.ignite.cdc.kafka.CdcKafkaReplicationTest.afterTest() timed out, dumping threads (afterTest() still keeps running)

As we can see in thread dump, method awaits topic deletion at line #92 of CdcKafkaReplicationTest.java [1]:

[08:13:22] : [Step 3/3] Thread [name="main", id=1, state=TIMED_WAITING, blockCnt=460, waitCnt=3547]
[08:13:22] : [Step 3/3] at java.lang.Thread.sleep(Native Method)
[08:13:22] : [Step 3/3] at o.a.i.i.util.IgniteUtils.sleep(IgniteUtils.java:8266)
[08:13:22] : [Step 3/3] at o.a.i.testframework.GridTestUtils.waitForCondition(GridTestUtils.java:2055)
[08:13:22] : [Step 3/3] at o.a.i.testframework.GridTestUtils.waitForCondition(GridTestUtils.java:2023)
[08:13:22] : [Step 3/3] at o.a.i.testframework.GridTestUtils.waitForCondition(GridTestUtils.java:2004)
[08:13:22] : [Step 3/3] at o.a.i.cdc.kafka.CdcKafkaReplicationTest.afterTest(CdcKafkaReplicationTest.java:92)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.GridAbstractTest.runAfterTest(GridAbstractTest.java:762)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.GridAbstractTest.cleanUpTestEnviroment(GridAbstractTest.java:739)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.GridAbstractTest.runTest(GridAbstractTest.java:2555)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.GridAbstractTest.access$600(GridAbstractTest.java:194)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.GridAbstractTest$1.evaluate(GridAbstractTest.java:237)
[08:13:22] : [Step 3/3] at org.junit.rules.TestWatcher$1.evaluate(TestWatcher.java:55)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.SystemPropertiesRule.lambda$methodStatement$1(SystemPropertiesRule.java:110)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.SystemPropertiesRule$$Lambda$56/739582925.evaluate(Unknown Source)
[08:13:22] : [Step 3/3] at o.a.i.testframework.junits.DelegatingJUnitStatement.evaluate(DelegatingJUnitStatement.java:49)

These timeouts leads to test execution over 5 minutes, as you can seen in [2-4] (sorted by test duration).
Reason should be investigated and fixed.

Links:

1. https://github.com/apache/ignite-extensions/blob/master/modules/cdc-ext/src/test/java/org/apache/ignite/cdc/kafka/CdcKafkaReplicationTest.java#L92
2. TC1 Build #412 at 13 Jan 08:02, build log: _TESTS_CDC_412.log.zip
3. TC1 Build #411 at 12 Jan 08:02, build log: _TESTS_CDC_411.log.zip
4. TC1 Build #407 at 10 Jan 08:00, build log: _TESTS_CDC_407.log.zip

Issue Links

• relates to: IGNITE-18209 Reduce binary metadata synchronization time for CDC through Kafka