[fix][client] Fix race condition between isDuplicate() and flushAsync() method in PersistentAcknowledgmentsGroupingTracker due to incorrect use Netty Recycler

[oneby-wang]

Fixes #25145

Motivation

See #25145 (comment).

Due to the use of Netty Recycler, a ConcurrentBitSetRecyclable instance may still be used after being recycled. However, recycling clears all bits, which may cause the isDuplicate() method return true under some race conditions.

If we execute to the point where bitSet != null, but the thread is then switched out and bitSet gets recycled, the isDuplicate() method will return true at that point.

pulsar/pulsar-client/src/main/java/org/apache/pulsar/client/impl/PersistentAcknowledgmentsGroupingTracker.java

Lines 135 to 138 in cab92f0

	if (messageIdAdv.getBatchIndex() >= 0) {
	ConcurrentBitSetRecyclable bitSet = pendingIndividualBatchIndexAcks.get(key);
	return bitSet != null && !bitSet.get(messageIdAdv.getBatchIndex());
	}

Modifications

Replace ConcurrentBitSetRecyclable with ConcurrentBitSet, ensure that ackSet is only cleared when the ackXxx method is invoked.

As #24766 suggested, we should stop using Netty Recycler in new code. Since using Netty Recycler is kind of error prone way here, I think we should remove it.

Verifying this change

• Make sure that the change passes the CI checks.

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

If the box was checked, please highlight the changes

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment

Documentation

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

Matching PR in forked repository

PR in forked repository: oneby-wang#26