[Bug] Failed to open ledger

[Technoboy-]

Search before asking

• I searched in the issues and found nothing similar.

Version

• Master branch.
• Branch-2.11.

Minimal reproduce step

Run test ReplicatorTest#testReplicationWithSchema in the master branch.

The test could run success if revert #17414.

What did you expect to see?

The test should pass

What did you see instead?

The test fails with :

org.apache.pulsar.broker.service.schema.exceptions.SchemaException: Error while recovering ledger -  ledger=2 - operation=Failed to open ledger
	at java.util.concurrent.CompletableFuture.encodeThrowable(CompletableFuture.java:332) ~[?:?]
	at java.util.concurrent.CompletableFuture.completeThrowable(CompletableFuture.java:347) ~[?:?]
	at java.util.concurrent.CompletableFuture$UniCompose.tryFire(CompletableFuture.java:1141) ~[?:?]
	at java.util.concurrent.CompletableFuture.postComplete(CompletableFuture.java:510) ~[?:?]
	at java.util.concurrent.CompletableFuture.completeExceptionally(CompletableFuture.java:2162) ~[?:?]

The root cause come from reading bk

Anything else?

No response

Are you willing to submit a PR?

• I'm willing to submit a PR!

[Technoboy-]

Need some help for this issue @zymap @horizonzy

[zymap]

I am taking a look

[zymap]

The root cause is this PR .

The problem is that we cannot call await() from an IO thread, even if we don’t have the consolidation in place
if the socket buffer is full, that blocking call would cause a deadlock, also without the consolidate
the await() was added, under the assumption that all these calls were coming from a thread pool dedicated to read operations.

I will push a fix to bookkeeper side and create a new release for fixing this.

[zymap]

Fixed through this PR apache/bookkeeper#3480

[github-actions]

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

[tisonkun]

Closed as fixed.