-
Notifications
You must be signed in to change notification settings - Fork 14.2k
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
MINOR: Speed up JoinGroupRequestTest, ListGroupsRequestTest, SyncGroupRequestTest by decreased rebalance deplay from 3s to 1s #16810
Conversation
@ExtendWith(value = Array(classOf[ClusterTestExtensions])) | ||
class JoinGroupRequestTest(cluster: ClusterInstance) extends GroupCoordinatorBaseRequestTest(cluster) { | ||
@ClusterTest(types = Array(Type.KRAFT), serverProperties = Array( | ||
new ClusterConfigProperty(key = "group.coordinator.new.enable", value = "true"), | ||
new ClusterConfigProperty(key = "offsets.topic.num.partitions", value = "1"), | ||
new ClusterConfigProperty(key = "offsets.topic.replication.factor", value = "1") | ||
new ClusterConfigProperty(key = "offsets.topic.replication.factor", value = "1"), | ||
new ClusterConfigProperty(key = "group.initial.rebalance.delay.ms", value = "0"), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Interesting finding, makes sense to me for the test. I expect it would it be helpful to add it also in the other test in this file, right?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@m1a2st Thanks for the patch. I understand the motivation but I wonder if it is a good choice. If we change the configuration, it means that the integration test won’t cover the initial rebalance delay which is the default behavior. I think that we should rather keep it as it is. |
Thanks for your comments, however I watch the kafka document
According to the document, thus I consider that these tests are suitable for this config. WDYT? |
The test sends many
Maybe we can reduce the delay from 3 seconds to 1.5 seconds? That not only keep the initial rebalance delay but also speed up the test. |
1230c8f
to
39e2dbb
Compare
@m1a2st Could you please share the "before" and "after"? |
According to the above test results, I consider we should modify |
39e2dbb
to
b25234f
Compare
that makes sense to me |
@m1a2st could you please revise the title? thanks! |
b25234f
to
9aedd6a
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@m1a2st thanks for updated PR
@@ -41,7 +41,7 @@ import scala.jdk.CollectionConverters._ | |||
* failure paths is to use timeouts. See {@link unit.kafka.server.BrokerRegistrationRequestTest} for integration test | |||
* of just the broker registration path. | |||
*/ | |||
@Timeout(120) | |||
@Timeout(30) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please revert this unrelated change
@@ -28,7 +28,7 @@ import org.junit.jupiter.api.Assertions.assertEquals | |||
import org.junit.jupiter.api.Timeout | |||
import org.junit.jupiter.api.extension.ExtendWith | |||
|
|||
@Timeout(120) | |||
@Timeout(30) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please revert this unrelated change
@@ -29,7 +29,7 @@ import org.junit.jupiter.api.extension.ExtendWith | |||
|
|||
import scala.jdk.CollectionConverters._ | |||
|
|||
@Timeout(120) | |||
@Timeout(30) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
please revert this unrelated change
9aedd6a
to
f36ac96
Compare
@chia7712, Thanks for your comments, revert it. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@m1a2st thanks for this patch. please attach the results again as the updated config is incorrect
@@ -51,7 +52,8 @@ class JoinGroupRequestTest(cluster: ClusterInstance) extends GroupCoordinatorBas | |||
@ClusterTest(serverProperties = Array( | |||
new ClusterConfigProperty(key = "group.coordinator.new.enable", value = "false"), | |||
new ClusterConfigProperty(key = "offsets.topic.num.partitions", value = "1"), | |||
new ClusterConfigProperty(key = "offsets.topic.replication.factor", value = "1") | |||
new ClusterConfigProperty(key = "offsets.topic.replication.factor", value = "1"), | |||
new ClusterConfigProperty(key = "group.initial.rebalance.delay.ms", value = "1"), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it should be "1000" rather than "1"
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes, the unit is ms, I fix it.
@m1a2st could you please share the results again? |
|
@m1a2st there are some tests getting slower. |
Becasue It is Zookeeper mode test, the config
|
makes sense. LGTM. will merge it later |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
@m1a2st could you please rebase code to trigger QA again? there are many failed tests. I feel they are unrelated, but double-check is not bad thing :) |
b77077a
to
d3bb650
Compare
I have rebased it :) |
@m1a2st please fix conflicts |
d3bb650
to
3f2d0e7
Compare
Thanks for your comments. I resolve it. |
With merging this PR we can keep testing the global timeout for all IT (#16957) |
…pRequestTest by decreased rebalance deplay from 3s to 1s (apache#16810) Reviewers: Chia-Ping Tsai <[email protected]>
…pRequestTest by decreased rebalance deplay from 3s to 1s (apache#16810) Reviewers: Chia-Ping Tsai <[email protected]>
Due to this test is too slow, it can't test in 30 seconds. Thus I add a config to speed up this test, let it can completed in 30s
Committer Checklist (excluded from commit message)