Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fixing Debezium integration tests #11154

Merged
merged 3 commits into from
Aug 4, 2021
Merged

Conversation

dlg99
Copy link
Contributor

@dlg99 dlg99 commented Jun 29, 2021

Fixes #11099

Motivation

Integration test for debezium connector to tun on CI

Modifications

Debezium integration test for postgres: added check for flush lsn updates

Added pulsar-io tests to the CI

Verifying this change

Ran locally:

mvn package -DintegrationTests -f tests/docker-images/pom.xml

mvn test -f tests/pom.xml test -DintegrationTestSuiteFile=pulsar-io-sources.xml -DintegrationTests -Dtest=org.apache.pulsar.tests.integration.io.sources.debezium.PulsarDebeziumSourcesTest
  • Make sure that the change passes the CI checks.

This change fixes existing tests.

Does this pull request potentially affect one of the following parts:

If yes was chosen, please highlight the changes

NO

Documentation

  • Does this pull request introduce a new feature? NO

@dlg99 dlg99 marked this pull request as draft June 29, 2021 21:13
@dlg99 dlg99 force-pushed the dbz-integration-tests branch 3 times, most recently from 0664a34 to 1cf310b Compare June 30, 2021 18:59
@dlg99 dlg99 marked this pull request as ready for review July 2, 2021 17:51
Copy link
Contributor

@eolivelli eolivelli left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Member

@lhotari lhotari left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Good work @dlg99 !

I was wondering how the tests weren't run in CI. I tracked it back to #10140 when tests were refactored. We missed the CI part for Pulsar IO in that PR. Thanks for fixing it as part of this PR.

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 7, 2021

/pulsarbot run-failure-checks

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 7, 2021

Pulsar IO integration test failed after rebase, likely the same reason as #11204 (comment)

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 7, 2021

I confirmed that tests pass after revert of PIP-85; @nlu90 is looking at the options to fix this without revert.

@eolivelli
Copy link
Contributor

Failed test
testDebeziumMongoDbSource(org.apache.pulsar.tests.integration.io.sources.debezium.PulsarDebeziumSourcesTest) Time elapsed: 33.666 s <<< FAILURE!
java.lang.AssertionError: expected [1] but found [0]
at org.testng.Assert.fail(Assert.java:99)
at org.testng.Assert.failNotEquals(Assert.java:1037)
at org.testng.Assert.assertEqualsImpl(Assert.java:140)
at org.testng.Assert.assertEquals(Assert.java:122)
at org.testng.Assert.assertEquals(Assert.java:907)
at org.testng.Assert.assertEquals

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 15, 2021

@nlu90 @eolivelli situation with this PR so far:
I rebased it on latest master and "CI - Integration - Pulsar-IO Sinks and Sources" failed.
Running testDebeziumMySqlSourceJson locally I can repro failures, rootcause TBD but the only suspicious thing that I see in the function log is

[pulsar-client-io-1-2] ERROR org.apache.pulsar.io.kafka.connect.AbstractKafkaConnectSource - Timed out waiting to flush org.apache.pulsar.io.kafka.connect.KafkaConnectSource$KafkaSourceRecord@72495bb1 offsets to storage

I reverted both PRs #11251 and #11056 locally; rebuilt/re-run the test and the test passed, "Timed out waiting to flush" is not logged in the function log

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 16, 2021

tests do pass if I create a new pulsar client in the PulsarOffsetBackingStore's ctor instead of the one passed there (I hardcoded the url in the test).

There is still something fishy with passing the client via source context and, ironically, I am missing on a context of what's going on there. Feels like some kind of deadlock at the PulsarOffsetStore.set() where it does producer.flushAsync().

I suggest we merge this PR, and @nlu90 continues work on the client in the SourceContext with the test enabled on the CI.

@nlu90
Copy link
Member

nlu90 commented Jul 19, 2021

@dlg99 Let me know when you merge this PR, I will continue look into the fix.

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 19, 2021

@nlu90 I don't have permissions to merge; I hope @codelipenghui / @sijie / @merlimat / @eolivelli or other committer can merge.

@eolivelli
Copy link
Contributor

@nlu90 are you "approving" this patch ?

@eolivelli
Copy link
Contributor

I am fine to merge this patch as soon as CI is green

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 20, 2021

@eolivelli

I am fine to merge this patch as soon as CI is green

I don't know what's going on with "CI - Unit - Brokers - Broker Group 1 / unit-tests" but it is not related.
"CI - Integration - Pulsar-IO Sinks and Sources" won't be green unless the problem with debezium/pulsar client in the source context is fixed.
The point of merging this to enable integration tests on the PR and break the circular dependency where @nlu90 does the fixes but I detect the problems by rebasing this PR.

@sijie
Copy link
Member

sijie commented Jul 23, 2021

I think the agreement is to let Neng work on a proper fix based on Andrey's PR.

@nlu90
Copy link
Member

nlu90 commented Jul 23, 2021

I pushed a fix for the debezium connector in #11435

And the integration test passed on my local laptop:

➜  sn-pulsar git:(neng/kafka-source-async-ack) ✗ mvn test -f tests/pom.xml test -DintegrationTestSuiteFile=pulsar-io-sources.xml -DintegrationTests -Dtest=org.apache.pulsar.tests.integration.io.sources.debezium.PulsarDebeziumSourcesTest
...
...
[INFO] ------------------------------------------------------------------------
[INFO] Reactor Summary for Apache Pulsar :: Tests 2.9.0-SNAPSHOT:
[INFO]
[INFO] Apache Pulsar :: Tests ............................. SUCCESS [  1.147 s]
[INFO] Apache Pulsar :: Tests :: Integration .............. SUCCESS [06:12 min]
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESS
[INFO] ------------------------------------------------------------------------
[INFO] Total time:  06:13 min
[INFO] Finished at: 2021-07-23T11:05:54-07:00
[INFO] ------------------------------------------------------------------------

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 23, 2021

@nlu90 I rebased the PR on top of recent master, the IO tests still failed.
The error looks different:

2021-07-23 21:25:17,576 - INFO  - [docker-java-stream-1663661242:DockerUtils$2@211] - DOCKER.exec(PulsarDebeziumSourcesTest-abzgk-pulsar-functions-worker-process-chgkv-1:/pulsar/bin/pulsar-admin source delete --tenant public --namespace default --name test-source-debezium-mysqlavro-PROCESS-mlijfzbu): Executing...
2021-07-23 21:25:21,122 - INFO  - [docker-java-stream-2004020930:DockerUtils$4@347] - DOCKER.exec(PulsarDebeziumSourcesTest-abzgk-pulsar-broker-0:tail -f /var/log/pulsar/broker.log): STDOUT: 21:25:21.000 [pulsar-web-40-6] INFO  org.eclipse.jetty.server.RequestLog - 172.21.0.9 - - [23/Jul/2021:21:25:20 +0000] "GET /admin/v2/persistent/public/functions/coordinate/stats?getPreciseBacklog=false&subscriptionBacklogSize=false HTTP/1.1" 200 2476 "-" "Pulsar-Java-v2.9.0-SNAPSHOT" 8
21:25:21.034 [pulsar-web-40-1] INFO  org.eclipse.jetty.server.RequestLog - 172.21.0.9 - - [23/Jul/2021:21:25:21 +0000] "GET /admin/v2/persistent/public/functions/coordinate/stats?getPreciseBacklog=false&subscriptionBacklogSize=false HTTP/1.1" 200 2476 "-" "Pulsar-Java-v2.9.0-SNAPSHOT" 16
./build/retry.sh: line 26: 31272 Killed                  "$@"
The command has failed after 3 attempts.

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 23, 2021

/pulsarbot run-failure-checks

@dlg99 dlg99 mentioned this pull request Jul 30, 2021
1 task
@dlg99
Copy link
Contributor Author

dlg99 commented Jul 30, 2021

/pulsarbot run-failure-checks

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 30, 2021

Excluded tests for sinks; something times out there, "Pulsar-IO Sinks and Sources" is limited to sources now (but passes)

@dlg99
Copy link
Contributor Author

dlg99 commented Jul 31, 2021

/pulsarbot run-failure-checks

@sijie sijie added this to the 2.9.0 milestone Aug 4, 2021
@sijie sijie merged commit 5d136a1 into apache:master Aug 4, 2021
LeBW pushed a commit to LeBW/pulsar that referenced this pull request Aug 9, 2021
Fixes apache#11099 

### Motivation

Integration test for debezium connector to tun on CI

### Modifications

Debezium integration test for postgres:  added check for flush lsn updates

Added pulsar-io tests to the CI
sijie pushed a commit that referenced this pull request Aug 13, 2021
Fixes #8502

### Motivation

Upgrade Debezium to a newer version

### Modifications

Upgraded Deebzium to v.1.5.4 (latest built with Java 8, v.1.6.x built with Java 11)
Upgraded kafka-client to 2.7 (version debezium tested with)
Scala-lib to 2.13.6 (for kafka-client)

Dealt with API changes, tests etc.

PR is on top of #11154 to have Debezium integration tests on CI.
@dlg99 dlg99 deleted the dbz-integration-tests branch October 14, 2021 23:31
bharanic-dev pushed a commit to bharanic-dev/pulsar that referenced this pull request Mar 18, 2022
Fixes apache#11099 

### Motivation

Integration test for debezium connector to tun on CI

### Modifications

Debezium integration test for postgres:  added check for flush lsn updates

Added pulsar-io tests to the CI
bharanic-dev pushed a commit to bharanic-dev/pulsar that referenced this pull request Mar 18, 2022
Fixes apache#8502

### Motivation

Upgrade Debezium to a newer version

### Modifications

Upgraded Deebzium to v.1.5.4 (latest built with Java 8, v.1.6.x built with Java 11)
Upgraded kafka-client to 2.7 (version debezium tested with)
Scala-lib to 2.13.6 (for kafka-client)

Dealt with API changes, tests etc.

PR is on top of apache#11154 to have Debezium integration tests on CI.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

PulsarDebeziumSourcesTest does not run on CI
5 participants