[Issue 8293][managed-ledger] Fix race condition in updating readPosition in ManagedCursorImpl

[lhotari]

Fixes #8293

Motivation

Regarding the race condition in #8229, it seems that ManagedCursorImpl.readPosition could get out of sync from OpReadEntry.readPosition if ManagedCursorImpl.readPosition gets updated after the OpReadEntry has been created since OpReadEntry's readPosition gets initialized from ManagedCursorImpl.readPosition.

The race condition seems to happen in this code in the setAcknowledgePosition method:

pulsar/managed-ledger/src/main/java/org/apache/bookkeeper/mledger/impl/ManagedCursorImpl.java

Lines 1512 to 1523 in 825fdd4

	if (readPosition.compareTo(newMarkDeletePosition) <= 0) {
	// If the position that is mark-deleted is past the read position, it
	// means that the client has skipped some entries. We need to move
	// read position forward
	PositionImpl oldReadPosition = readPosition;
	readPosition = ledger.getNextValidPosition(newMarkDeletePosition);
	if (log.isDebugEnabled()) {
	log.debug("[{}] Moved read position from: {} to: {}, and new mark-delete position {}", ledger.getName(),
	oldReadPosition, readPosition, markDeletePosition);
	}
	}

The intent of this change is to properly handle concurrent updates when updating readPosition in setAcknowledgePosition method.

Modifications

The update of readPosition uses READ_POSITION_UPDATER.updateAndGet to handle concurrent updates to readPosition in setAcknowledgePosition method.

Verifying this change

There are currently no separate tests to verify this change.

[lhotari]

/pulsarbot run-failure-checks

[lhotari]

/pulsarbot run-failure-checks

[merlimat]

👍 Nice catch!