lhotari edited a comment on issue #10832: URL: https://github.com/apache/pulsar/issues/10832#issuecomment-856716926
> What if an async sent message failed, and for example, say, it's the N/2-th message, and I retry sent this message, eventually, does the consumer still can read those messages in order? Is there any chance to cause messages disorder? I believe that with `.blockIfQueueFull(true)` the chances for ordering problems is not a big concern. However there might not be a guarantee that message order is preserved if the results of sendAsync aren't checked before sending the next message. It's worth checking with tests what your experiences are. With the default setting of `.blockIfQueueFull(false)` and using `sendAsync`, there's a real challenge how to handle the queue full exceptions which are used as a flow control method. > And why we need to ensure after flush()? you mean it flushes all backed in-memory messages safely persisted in Pulsar broker? and for those failed messages, it won't in the buffered to-be-sent list, so you have to retry for it? Flush can succeed regardless of any possible previous failures in message sending. The only way to ensure that the messages have been sent is to check the results. Checking the CompletableFuture instances could happen after calling flush(). Perhaps there are better ways to ensure message delivery with batches when using transactions with Pulsar. However I haven't tried that myself. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected]
