acknowledgmentAtBatchIndexLevelEnabled will still receive duplicate messages when unload topic

[xuesongxs]

Describe the bug Pulsar v2.8.3 acknowledgmentAtBatchIndexLevelEnabled will still receive duplicate messages

To Reproduce Steps to reproduce the behavior:

1. Set acknowledgmentAtBatchIndexLevelEnabled=true in broker.conf

2. Create consumer:

   public static void main(String[] args) throws Exception{
       PulsarClient client = PulsarClient
               .builder()
               .serviceUrl("pulsar://127.0.0.1:6650")
               .build();
   
       ConsumerBuilder<NlMessage> consumerBuilder = client.newConsumer(Schema.JSON(NlMessage.class)).topic("batchtest")
               .subscriptionName("batchTest")
               .subscriptionType(SubscriptionType.Shared)
               .enableRetry(true)
               .enableBatchIndexAcknowledgment(true)
               .receiverQueueSize(1000)
               .negativeAckRedeliveryDelay(20, TimeUnit.SECONDS)
               .subscriptionInitialPosition(SubscriptionInitialPosition.Earliest);
       // DLQ
       DeadLetterPolicy deadLetterPolicy = DeadLetterPolicy.builder()
               .deadLetterTopic("persistent://public/default/batchtest-dlq")
               .retryLetterTopic("persistent://public/default/batchtest-retry")
               .maxRedeliverCount(3)
               .build();
       consumerBuilder.deadLetterPolicy(deadLetterPolicy);
       consumerBuilder.negativeAckRedeliveryDelay(10, TimeUnit.MICROSECONDS);
       consumerBuilder.ackTimeout(0, TimeUnit.SECONDS);
   
       Consumer<NlMessage> consumer = consumerBuilder.subscribe();
       long i = 0;
       while (true) {
           Message<NlMessage> message = consumer.receive();
           System.out.println("message:" + message.getMessageId());
           if (i <= 99) {
               consumer.acknowledge(message);
               System.out.println("acked message:" + i);
           } else {
               // hold the msg
           }
           i++;
       }

3. Create producer and send 100 msgs:

   public static void main(String[] args) throws PulsarClientException {
       PulsarClient pulsarClient = PulsarClient.builder().serviceUrl("pulsar://127.0.0.1:6650").build();
       Producer<NlMessage> producer = pulsarClient.newProducer(Schema.JSON(NlMessage.class))
               .topic("persistent://public/default/batchtest")
               .blockIfQueueFull(true)
               .sendTimeout(0, TimeUnit.SECONDS)
               .create();
       long i = 1;
       while(i <= 100L){
           producer.sendAsync(new NlMessage());
           i++;
       }
   }

4. 100 msgs in 1 backlog
5. After 2 minutes, execute unload cmd: bin/pulsar-admin topics unload persistent://public/default/batchtest
6. See consumer log, there will also be several or more messages that have been ack pushed.

Expected behavior Only push the message without ack.

Screenshots If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: [e.g. iOS]

Additional context Add any other context about the problem here.

[Jason918]

Is it reproduced on master branch? Do you plan to work on a fix? @xuesongxs

[xuesongxs]

Is it reproduced on master branch? Do you plan to work on a fix? @xuesongxs

I tested it on the 2.8.3 branch, the same should be true for the master branch, I won't fix this.

[Technoboy-]

In this case, it may receive the duplicate message.

[Jason918]

In this case, it may receive the duplicate message.

Because mark delete info is persistent periodically and it's not triggered when topic unloads?

[xuesongxs]

In this case, it may receive the duplicate message.

Because mark delete info is persistent periodically and it's not triggered when topic unloads?

I unloaded or stopped the broker 5 minutes after the ack of the first 99 messages was completed. Even if I write index asynchronously, 5 minutes should be enough, but there is still a problem.

[lhotari]

15031 and #15067 fix issues in this area.

[github-actions[bot]]

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

[github-actions[bot]]

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

[xuesongxs]

15031 and #15067 fix issues in this area.

I have verified these 2 PRs in the latest pulsar version and have not solved this bug.

[github-actions[bot]]

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

[jdfrozen]

2.8.4 duplicated delayed messages when unload topic