Page History

public abstract class AsyncSinkWriter<InputT, RequestEntryT extends Serializable> implements SinkWriter<InputT, AtomicLongVoid, Collection<RequestEntryT>> {

    /**
     * This method specifies how to persist buffered request entries into the
     * destination. It is implemented when support for a new destination is
     * added.
     * <p>
     * The method is invoked with a set of request entries according to the
     * buffering hints (and the valid limits of the destination). The logic then
     * needs to create and execute the request against the destination (ideally
     * by batching together multiple request entries to increase efficiency).
     * The logic also needs to identify individual request entries that were not
     * persisted successfully and resubmit them using the {@code
     * requeueFailedRequestEntryrequestResult} method.
     * <p>
     * During checkpointing, the sink needs to ensure that there are no
     * outstanding in-flight requests.
     *
     * @param requestEntries a set of request entries that should be sent to the
     *                       destination
     * @param requestResult  a ResultFuture that needs to be completed once all
     *                       request entries that have been passed to the method
     *                       on invocation have either been successfully
     *                       persisted in the destination or have been
     *                       re-queued
     */
    protected abstract void submitRequestEntries(List<RequestEntryT> requestEntries, ResultFuture<?>ResultFuture<RequestEntryT> requestResult);

    ...
}

    /**
     * The ElementConverter provides a mapping between for the elements of a
     * stream to request entries that can be sent to the destination.
     * <p>
     * The resulting request entry is buffered by the AsyncSinkWriter and sent
     * to the destination when the {@code submitRequestEntries} method is
     * invoked.
     */
    private final ElementConverter<InputT, RequestEntryT> elementConverter;


    /**
     * Buffer to hold request entries that should be persisted into the
     * destination.
     * <p>
     * A request entry contain all relevant details to make a call to the
     * destination. Eg, for Kinesis Data Streams a request entry contains the
     * payload and partition key.
     * <p>
     * It seems more natural to buffer InputT, ie, the events that should be
     * persisted, rather than RequestEntryT. However, in practice, the response
     * of a failed request call can make it very hard, if not impossible, to
     * reconstruct the original event. It is much easier, to just construct a
     * new (retry) request entry from the response and add that back to the
     * queue for later retry.
     */
    private final BlockingDeque<RequestEntryT>Deque<RequestEntryT> bufferedRequestEntries = new LinkedBlockingDeque<>ArrayDeque<>(...);


    /**
     * Tracks all pending async calls that have been executed since the last
     * checkpoint. Calls that completed (successfully or unsuccessfully) are
     * automatically decrementing the counter. Any request entry that was not
     * successfully persisted needs to be handled and retried by the logic in
     * {@code submitRequestsToApi}.
     * <p>
     * There is a limit on the number of concurrent (async) requests that can be
     * handled by the client library. This limit is enforced by checking the
     * size of this queue before issuing new requests.
     * <p>
     * To complete a checkpoint, we need to make sure that no requests are in
     * flight, as they may fail, which could then lead to data loss.
     */
    private AtomicLongint numberOfInFlightRequestsinFlightRequestsCount;


    @Override
    public void write(InputT element, Context context) throws IOException, InterruptedException {
        // blocks if too many eventselements have been buffered
        while (bufferedRequestEntries.putLast(elementConverter.apply(element, context));

   size() >= MAX_BUFFERED_REQUESTS_ENTRIES) {
     // blocks if too many async requests are in flight mailboxExecutor.yield();
        if (/*buffering hints are met*/) {
			}

        bufferedRequestEntries.add(elementConverter.apply(element, context));

        // blocks if too many async requests are in flight
        flush();
		}
    }


    /**
     * The entire request may fail or single request entries that are part of
     * the request may not be persisted successfully, eg, because of network
     * issues or service side throttling. All request entries that failed with
     * transient failures need to be re-queued with this method so that aren't
     * lost and can be retried later.
     * <p>
     * Request entries that are causing the same error in a reproducible manner,
     * eg, ill-formed request entries, must not be re-queued but the error needs
     * to be handled in the logic of {@code submitRequestEntries}. Otherwise
     * these request entries will be retried indefinitely, always causing the
     * same error.
     */
    protected void requeueFailedRequestEntry(RequestEntryT requestEntry) throws InterruptedException {
        bufferedRequestEntries.putFirst(requestEntry);
    }


    /**
     * Persists buffered RequestsEntries into the destination by invoking {@code
     * submitRequestEntries} with batches according to the user specified
     * buffering hints.
     *
     * The method blocks if too many async requests are in flight.
     */
    private void flush() throws InterruptedException {
        while (bufferedRequestEntries.size() >= MAX_BATCH_SIZE) {
Persists buffered RequestsEntries into the destination by invoking {@code
     * submitRequestEntries} with batches according to the user specified
     * buffering hints.
     *
     * The method blocks if too many async requests are in flight.
     */
    private void flush() throws InterruptedException {
        while (bufferedRequestEntries.size() >= MAX_BATCH_SIZE) {

            // create a batch of request entries that should be persisted in the destination
            ArrayList<RequestEntryT> batch = new ArrayList<>(MAX_BATCH_SIZE);

            while (batch.size() <= MAX_BATCH_SIZE && !bufferedRequestEntries.isEmpty()) {
                try {
                    batch.add(bufferedRequestEntries.remove());
                } catch (NoSuchElementException e) {
                    // if there are not enough elements, just create a smaller batch
                    break;
                }
            }

            ResultFuture<RequestEntryT> requestResult =
                    failedRequestEntries -> mailboxExecutor.execute(
                            () -> completeRequest(failedRequestEntries),
                            "Mark in-flight request as completed and requeue %d request entries",
               // create a batch of request entries that should be persisted in the destination failedRequestEntries.size());

            ArrayList<RequestEntryT>while batch(inFlightRequestsCount >= new ArrayList<>(MAX_BATCHIN_FLIGHT_SIZEREQUESTS); {

            for (int i = 0; i < MAX_BATCH_SIZE; i++) {mailboxExecutor.yield();
            }

    try {
       inFlightRequestsCount++;
             batch.add(bufferedRequestEntries.remove()submitRequestEntries(batch, requestResult);
        }
        } catch (NoSuchElementException e) {}


    /**
     * Marks an in-flight request as completed and prepends failed requestEntries back to the
   // if there* areinternal notrequestEntry enoughbuffer elements,for justlater createretry.
 a smaller batch
  *
     * @param failedRequestEntries requestEntries that need to be retried
     break;*/
    private void completeRequest(Collection<RequestEntryT> failedRequestEntries) {
         }inFlightRequestsCount--;

        // By just  }

			ResultFuture<?> requestResult = ...

    iterating through failedRequestEntries, it reverses the order of the
        if// (numberOfInFlightRequestsfailedRequestEntries.getAndIncrement() >= MAX_IN_FLIGHT_REQUESTS) {
        It doesn't make a difference for kinesis:putRecords, as the api
        // blockdoes andnot waitmake untilany enough in-fligh requests have completed
    order guarantees, but may cause avoidable reorderings for other
        }

			// call the destination specific code that actually persists the request entriesdestinations.
            submitRequestEntries(batch, requestResultfailedRequestEntries.forEach(bufferedRequestEntries::addFirst);
    }


    /**
     * In flight requests will be retried if the sink is still healthy. But if
     * in-flight requests fail after a checkpoint has been triggered and Flink
     * needs to recover from the checkpoint, the (failed) in-flight requests are
     * gone and cannot be retried. Hence, there cannot be any outstanding
     * in-flight requests when a commit is initialized.
     * <p>
     * To this end, all in-flight requests need to be passed to the {@code
     * AsyncSinkCommiter} in order to be completed as part of the pre commit.
     */
    @Override
    public List<AtomicLong> prepareCommit(boolean flush) throws IOException {
        logger.info("Prepare commit. {} requests currently in flight.", numberOfInFlightRequests.get());

        // reuse current inFlightRequests as commitable and create empty queue to avoid copy and clearing
        List<AtomicLong> committable = Collections.singletonList(numberOfInFlightRequests);

        // all in-flight requests are handled by the AsyncSinkCommiter and new elements cannot be added to the queue during a commit, so it's save to create a counter
        numberOfInFlightRequests = new AtomicLong();

        return committable;
    }


    /**
     * All in-flight requests have been completed, but there may still be
     * request entries in the internal buffer that are yet to be sent to the
     * endpoint. These request entries are stored in the snapshot state so that
     * they don't get lost in case of a failure/restart of the application.
     */
    @Override
    public List<Collection<RequestEntryT>> snapshotState() throws IOException {
        return Collections.singletonList(bufferedRequestEntries);
    }