[fix][broker] Remove failed future when topic load error #20561

poorbarcode · 2023-06-12T11:16:35Z

Motivation

Issue-1(this PR fixes)

When a topic load fails due to timeout ex, Broker will not clean up the exception future, causing other APIs to fail when they touch this future. You can reproduce it by testUnloadNamespaceAfterLoadTopicFailed

Issue-2

If the program is executed as the flow below, there will still be one orphan topic that is not closed.

time	`create topic`	`timeout check`
1	Start to create a topic
2	Open managed ledger, and do recover for every cursor
3		Time out and remove the topic from the cache
4	An orphan topic still is there

I will try to reproduce this scenario later(not in this PR)

Contexts

2023-06-12T06:55:17,479+0000 [bookkeeper-ml-scheduler-OrderedScheduler-21-0] ERROR
	at java.util.concurrent.CompletableFuture.encodeThrowable(CompletableFuture.java:331) ~[?:?]
	at java.util.concurrent.CompletableFuture.uniComposeStage(CompletableFuture.java:1099) ~[?:?]
	at java.util.concurrent.CompletableFuture.thenCompose(CompletableFuture.java:2235) ~[?:?]
	at org.apache.pulsar.broker.service.BrokerService.lambda$unloadServiceUnit$88(BrokerService.java:1951) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.common.util.collections.ConcurrentOpenHashMap$Section.forEach(ConcurrentOpenHashMap.java:544) ~[io.streamnative-pulsar-common-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.common.util.collections.ConcurrentOpenHashMap.forEach(ConcurrentOpenHashMap.java:272) ~[io.streamnative-pulsar-common-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.service.BrokerService.unloadServiceUnit(BrokerService.java:1945) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.service.BrokerService.unloadServiceUnit(BrokerService.java:1921) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.namespace.OwnedBundle.lambda$handleUnloadRequest$0(OwnedBundle.java:138) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at java.util.concurrent.CompletableFuture.uniComposeStage(CompletableFuture.java:1106) ~[?:?]
	at java.util.concurrent.CompletableFuture.thenCompose(CompletableFuture.java:2235) ~[?:?]
	at org.apache.pulsar.broker.namespace.OwnedBundle.handleUnloadRequest(OwnedBundle.java:138) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.namespace.NamespaceService.unloadNamespaceBundle(NamespaceService.java:734) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.namespace.NamespaceService.unloadNamespaceBundle(NamespaceService.java:723) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.namespace.NamespaceService.unloadNamespaceBundle(NamespaceService.java:719) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.admin.impl.NamespacesBase.lambda$internalUnloadNamespaceBundle$51(NamespacesBase.java:1228) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at java.util.concurrent.CompletableFuture.uniAcceptNow(CompletableFuture.java:753) ~[?:?]
	at java.util.concurrent.CompletableFuture.uniAcceptStage(CompletableFuture.java:731) ~[?:?]
	at java.util.concurrent.CompletableFuture.thenAccept(CompletableFuture.java:2108) ~[?:?]
	at org.apache.pulsar.broker.admin.impl.NamespacesBase.internalUnloadNamespaceBundle(NamespacesBase.java:1211) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at org.apache.pulsar.broker.admin.v2.Namespaces.unloadNamespaceBundle(Namespaces.java:555) ~[io.streamnative-pulsar-broker-2.10.3.5.jar:2.10.3.5]
	at jdk.internal.reflect.GeneratedMethodAccessor131.invoke(Unknown Source) ~[?:?]
	at jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) ~[?:?]
	at java.lang.reflect.Method.invoke(Method.java:566) ~[?:?]
	at org.glassfish.jersey.server.model.internal.ResourceMethodInvocationHandlerFactory.lambda$static$0(ResourceMethodInvocationHandlerFactory.java:52) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.internal.AbstractJavaResourceMethodDispatcher$1.run(AbstractJavaResourceMethodDispatcher.java:124) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.internal.AbstractJavaResourceMethodDispatcher.invoke(AbstractJavaResourceMethodDispatcher.java:167) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.internal.JavaResourceMethodDispatcherProvider$VoidOutInvoker.doDispatch(JavaResourceMethodDispatcherProvider.java:159) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.internal.AbstractJavaResourceMethodDispatcher.dispatch(AbstractJavaResourceMethodDispatcher.java:79) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.ResourceMethodInvoker.invoke(ResourceMethodInvoker.java:475) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.ResourceMethodInvoker.apply(ResourceMethodInvoker.java:397) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.model.ResourceMethodInvoker.apply(ResourceMethodInvoker.java:81) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.server.ServerRuntime$1.run(ServerRuntime.java:255) ~[org.glassfish.jersey.core-jersey-server-2.34.jar:?]
	at org.glassfish.jersey.internal.Errors$1.call(Errors.java:248) ~[org.glassfish.jersey.core-jersey-common-2.34.jar:?]
	at org.glassfish.jersey.internal.Errors$1.call(Errors.java:244) ~[org.glassfish.jersey.core-jersey-common-2.34.jar:?]
	at org.glassfish.jersey.internal.Errors.process(Errors.java:292) ~[org.glassfish.jersey.core-jersey-common-2.34.jar:?]
	at org.glassfish.jersey.internal.Errors.process(Errors.java:274) ~[org.glassfish.jersey.core-jersey-common-2.34.jar:?]
	at org.glassfish.jersey.internal.Errors.process(Errors.java:244) ~[org.glassfish.jersey.core-jersey-common-2.34.jar:?]

Modifications

Remove failed future when topic load error.

Documentation

doc
doc-required
doc-not-needed
doc-complete

Matching PR in forked repository

PR in forked repository: x

eolivelli · 2023-06-12T12:31:47Z

pulsar-broker/src/main/java/org/apache/pulsar/broker/service/BrokerService.java

                });
            }
+            final CompletableFuture<Optional<Topic>> topicFutureToRemove = topicFuture;
+            topicFuture.whenComplete((ignore, ex) -> topics.remove(topicName.toString(), topicFutureToRemove));


please note that you are appending a new whenComplete listener every time this method is called and the object is already in the map.
we have to call "whenComplete" only if the body of the lambda expression in computeIfAbsent is executed.
I know that it sounds ugly, but we must to it

@eolivelli

Ah, sorry. I lost a check: if (ex != null). It only is removed if the future was failed, and we call remove(k, v), and this method will not do incorrect deletions. Could you take a look again?

The problem is that we call whenComplete every time we invoke this method and only the first time we create the object and add it to the Map.

So everytime we enqueue the execution. If this method is called in an hotpath you will see many allocations

@eolivelli

You are right. Thanks, already fixed

lhotari · 2023-07-05T16:49:22Z

this is already covered in #20540 which was created before this PR. It uses a different approach in the failTopicFuture and completeTopicFuture method.
I agree that the approach in this PR could resolve the particular issue, but #20540 targets to cover also the cases where the bundle gets unloaded or the broker gets shutdown while a topic is loading.

github-actions · 2023-08-31T01:47:37Z

The pr had no activity for 30 days, mark with Stale label.

poorbarcode added 3 commits June 12, 2023 18:23

[fix][broker] Remove timeout cached topic after topic load error

e61d200

[fix][broker] Remove timeout cached topic after topic load error

93037c1

[fix][broker] Remove timeout cached topic after topic load error

ab36d47

github-actions bot added the doc-not-needed Your PR changes do not impact docs label Jun 12, 2023

poorbarcode self-assigned this Jun 12, 2023

poorbarcode added release/2.10.5 release/2.11.2 release/3.0.1 release/3.0.2 and removed release/3.0.1 labels Jun 12, 2023

poorbarcode added this to the 3.1.0 milestone Jun 12, 2023

eolivelli requested changes Jun 12, 2023

View reviewed changes

Avoid deleting successful future

2e414de

poorbarcode mentioned this pull request Jun 13, 2023

[fix][broker] Fix multiple race conditions in topic unloading and loading #20540

Open

4 tasks

append action only new future is creting

4f2c001

poorbarcode requested a review from eolivelli June 13, 2023 07:45

Technoboy- modified the milestones: 3.1.0, 3.2.0 Jul 31, 2023

github-actions bot added the Stale label Aug 31, 2023

poorbarcode added release/3.0.3 and removed release/3.0.2 labels Oct 24, 2023

Technoboy- modified the milestones: 3.2.0, 3.3.0 Dec 22, 2023

coderzc modified the milestones: 3.3.0, 3.4.0 May 8, 2024

lhotari added release/3.0.8 and removed release/3.0.3 labels Oct 10, 2024

lhotari removed this from the 4.0.0 milestone Oct 14, 2024

lhotari added this to the 4.1.0 milestone Oct 14, 2024

lhotari added release/3.0.9 and removed release/3.0.8 labels Dec 6, 2024

lhotari added release/3.0.10 and removed release/3.0.9 release/2.10.5 release/2.11.2 release/3.0.10 labels Jan 3, 2025

lhotari removed this from the 4.1.0 milestone Jan 6, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[fix][broker] Remove failed future when topic load error #20561

[fix][broker] Remove failed future when topic load error #20561

Uh oh!

poorbarcode commented Jun 12, 2023 •

edited

Loading

Uh oh!

eolivelli Jun 12, 2023

Uh oh!

poorbarcode Jun 12, 2023

Uh oh!

eolivelli Jun 12, 2023

Uh oh!

poorbarcode Jun 13, 2023

Uh oh!

lhotari commented Jul 5, 2023

Uh oh!

github-actions bot commented Aug 31, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

[fix][broker] Remove failed future when topic load error #20561

Are you sure you want to change the base?

[fix][broker] Remove failed future when topic load error #20561

Uh oh!

Conversation

poorbarcode commented Jun 12, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Motivation

Issue-1(this PR fixes)

Issue-2

Contexts

Modifications

Documentation

Matching PR in forked repository

Uh oh!

eolivelli Jun 12, 2023

Choose a reason for hiding this comment

Uh oh!

poorbarcode Jun 12, 2023

Choose a reason for hiding this comment

Uh oh!

eolivelli Jun 12, 2023

Choose a reason for hiding this comment

Uh oh!

poorbarcode Jun 13, 2023

Choose a reason for hiding this comment

Uh oh!

lhotari commented Jul 5, 2023

Uh oh!

github-actions bot commented Aug 31, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

poorbarcode commented Jun 12, 2023 •

edited

Loading