rkhachatryan commented on a change in pull request #11515: [FLINK-16744][task]
implement channel state persistence for unaligned checkpoints
URL: https://github.com/apache/flink/pull/11515#discussion_r401885233
##########
File path:
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/SubtaskCheckpointCoordinatorImpl.java
##########
@@ -89,17 +101,98 @@ public void checkpointState(
checkpointMetaData.getCheckpointId(),
checkpointOptions.getTargetLocation());
- CheckpointingOperation.execute(
- checkpointMetaData,
- checkpointOptions,
- checkpointMetrics,
- storage,
- operatorChain,
- taskName,
- closeableRegistry,
- executorService,
- env,
- asyncExceptionHandler,
- isCanceled);
+ long startSyncPartNano = System.nanoTime();
+
+ HashMap<OperatorID, OperatorSnapshotFutures>
operatorSnapshotsInProgress = new
HashMap<>(operatorChain.getNumberOfOperators());
+ try {
+ for (StreamOperatorWrapper<?, ?> operatorWrapper :
operatorChain.getAllOperators(true)) {
+ StreamOperator<?> op =
operatorWrapper.getStreamOperator();
+ OperatorSnapshotFutures snapshotInProgress =
checkpointStreamOperator(
+ op,
+ checkpointMetaData,
+ checkpointOptions,
+ storage,
+ isCanceled);
+
operatorSnapshotsInProgress.put(op.getOperatorID(), snapshotInProgress);
+ }
+
+ if (LOG.isDebugEnabled()) {
+ LOG.debug("Finished synchronous checkpoints for
checkpoint {} on task {}",
+ checkpointMetaData.getCheckpointId(),
taskName);
+ }
+
+ long startAsyncPartNano = System.nanoTime();
+
+
checkpointMetrics.setSyncDurationMillis((startAsyncPartNano -
startSyncPartNano) / 1_000_000);
+
+ // we are transferring ownership over
snapshotInProgressList for cleanup to the thread, active on submit
+ executorService.execute(new AsyncCheckpointRunnable(
+ operatorSnapshotsInProgress,
+ checkpointMetaData,
+ checkpointMetrics,
+ startAsyncPartNano,
+ taskName,
+ closeableRegistry,
+ env,
+ asyncExceptionHandler));
+
+ if (LOG.isDebugEnabled()) {
+ LOG.debug(
+ "{} - finished synchronous part of
checkpoint {}. Alignment duration: {} ms, snapshot duration {} ms",
+ taskName,
checkpointMetaData.getCheckpointId(),
+
checkpointMetrics.getAlignmentDurationNanos() / 1_000_000,
+
checkpointMetrics.getSyncDurationMillis());
+ }
+ } catch (Exception ex) {
+ // Cleanup to release resources
+ for (OperatorSnapshotFutures operatorSnapshotResult :
operatorSnapshotsInProgress.values()) {
+ if (null != operatorSnapshotResult) {
+ try {
+ operatorSnapshotResult.cancel();
+ } catch (Exception e) {
+ LOG.warn("Could not properly
cancel an operator snapshot result.", e);
+ }
+ }
+ }
+
+ if (LOG.isDebugEnabled()) {
+ LOG.debug(
+ "{} - did NOT finish synchronous part
of checkpoint {}. Alignment duration: {} ms, snapshot duration {} ms",
+ taskName,
checkpointMetaData.getCheckpointId(),
+
checkpointMetrics.getAlignmentDurationNanos() / 1_000_000,
+
checkpointMetrics.getSyncDurationMillis());
+ }
+
+ if
(checkpointOptions.getCheckpointType().isSynchronous()) {
+ // in the case of a synchronous checkpoint, we
always rethrow the exception,
+ // so that the task fails.
+ // this is because the intention is always to
stop the job after this checkpointing
+ // operation, and without the failure, the task
would go back to normal execution.
+ throw ex;
+ } else {
+
env.declineCheckpoint(checkpointMetaData.getCheckpointId(), ex);
+ }
+ }
+ }
+
+ private static OperatorSnapshotFutures checkpointStreamOperator(
+ StreamOperator<?> op,
+ CheckpointMetaData checkpointMetaData,
+ CheckpointOptions checkpointOptions,
+ CheckpointStreamFactory storageLocation,
+ Supplier<Boolean> isCanceled) throws Exception {
+ try {
+ return op.snapshotState(
+ checkpointMetaData.getCheckpointId(),
+ checkpointMetaData.getTimestamp(),
+ checkpointOptions,
+ storageLocation);
+ }
+ catch (Exception ex) {
+ if (!isCanceled.get()) {
+ LOG.info(ex.getMessage(), ex);
Review comment:
This change was introduced by another commit into `CheckpointingOperation`
which I merged with `SubtaskCheckpointCoordinatorImpl`.
As I see from `StreamOperatorStateHandler.snapshotState`, the message is
built like this:
```
String snapshotFailMessage = "Could not complete snapshot " + checkpointId +
" for operator " + operatorName + ".";
```
Which I think provides enough details.
Not sure, why do we log (info) and re-throw it, but I'd like not to address
it in this PR.
----------------------------------------------------------------
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]
With regards,
Apache Git Services
