Skip to content

[fix][broker] Fix update topic partitions failed #15217

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.

Sign up for GitHub

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

Closed
TakaHiR07 wants to merge 2 commits into from
Closed

[fix][broker] Fix update topic partitions failed #15217

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3d554d1
After catch the "Subscription already exists" error, do updatePartiti…
Apr 19, 2022
e90def9
no need to clean up znode when updatePartitionedMetadata failed
Apr 21, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 8 additions & 22 deletions pulsar-broker/src/main/java/org/apache/pulsar/broker/admin/impl/PersistentTopicsBase.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4029,31 +4029,17 @@ private PersistentReplicator getReplicatorReference(String replName, PersistentT

private CompletableFuture<Void> updatePartitionedTopic(TopicName topicName, int numPartitions, boolean force) {
CompletableFuture<Void> result = new CompletableFuture<>();
createSubscriptions(topicName, numPartitions).thenCompose(__ -> {
CompletableFuture<Void> future = namespaceResources().getPartitionedTopicResources()
.updatePartitionedTopicAsync(topicName, p -> new PartitionedTopicMetadata(numPartitions));
future.exceptionally(ex -> {
// If the update operation fails, clean up the partitions that were created
getPartitionedTopicMetadataAsync(topicName, false, false).thenAccept(metadata -> {
int oldPartition = metadata.partitions;
for (int i = oldPartition; i < numPartitions; i++) {
topicResources().deletePersistentTopicAsync(topicName.getPartition(i)).exceptionally(ex1 -> {
log.warn("[{}] Failed to clean up managedLedger {}", clientAppId(), topicName,
ex1.getCause());
return null;
});
}
}).exceptionally(e -> {
log.warn("[{}] Failed to clean up managedLedger", topicName, e);
createSubscriptions(topicName, numPartitions).thenCompose(__ -> namespaceResources().getPartitionedTopicResources()
.updatePartitionedTopicAsync(topicName, p -> new PartitionedTopicMetadata(numPartitions))
).thenAccept(__ -> result.complete(null)).exceptionally(ex -> {
if (force && ex.getCause() instanceof PulsarAdminException.ConflictException) {
CompletableFuture<Void> future = namespaceResources().getPartitionedTopicResources()
.updatePartitionedTopicAsync(topicName, p -> new PartitionedTopicMetadata(numPartitions));
future.thenAccept(__ -> result.complete(null)).exceptionally(ex1 -> {
result.completeExceptionally(ex1);
return null;
});

@Technoboy- Technoboy- Apr 19, 2022

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

The updatePartitionedTopicAsync has handled the exception, so it can't delegate the exception here.
BTW, could you help refactor this method ?

@TakaHiR07 TakaHiR07 Apr 19, 2022

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

I think it do not handle since the code between 4033-4035 also delegate the exception. And I think I can help refactor this method.

@gaozhangmin gaozhangmin Apr 20, 2022 •
edited
Loading

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

@takahiro0208 If createSubscriptions failed, there would be a situation which sub Znode under managed-ledgers had created. your deletion here will fail also.

@TakaHiR07 TakaHiR07 Apr 20, 2022

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

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

@takahiro0208 If createSubscriptions failed, there would be a situation which sub Znode under managed-ledgers had created. your deletion here will fail also.

If we use "force=true" to retry updatePartition, it seems no need to clean-up znode ?

return null;
});
return future;
}).thenAccept(__ -> result.complete(null)).exceptionally(ex -> {
if (force && ex.getCause() instanceof PulsarAdminException.ConflictException) {
result.complete(null);
return null;
}
result.completeExceptionally(ex);
return null;
Expand Down
8 changes: 7 additions & 1 deletion pulsar-broker/src/test/java/org/apache/pulsar/broker/admin/AdminApi2Test.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2350,7 +2350,13 @@ public void testFailedUpdatePartitionedTopic() throws Exception {
}
admin.topics().updatePartitionedTopic(partitionedTopicName, newPartitions, false, true);
// validate subscription is created for new partition.
assertNotNull(admin.topics().getStats(partitionedTopicName + "-partition-" + 6).getSubscriptions().get(subName1));
for (int i = startPartitions; i < newPartitions; i++) {
assertNotNull(
admin.topics().getStats(partitionedTopicName + "-partition-" + i).getSubscriptions().get(subName1));
}

// validate update partition is success
assertEquals(admin.topics().getPartitionedTopicMetadata(partitionedTopicName).partitions, newPartitions);
}

@Test(dataProvider = "topicType")
Expand Down