Github user tdas commented on a diff in the pull request:
https://github.com/apache/spark/pull/7825#discussion_r36041342
--- Diff:
extras/kinesis-asl/src/main/scala/org/apache/spark/streaming/kinesis/KinesisReceiver.scala
---
@@ -182,4 +194,82 @@ private[kinesis] class KinesisReceiver(
new DefaultAWSCredentialsProviderChain()
}
}
+
+ /** Remember the range of sequence numbers that was added to the
currently active block */
+ private def rememberAddedRange(range: SequenceNumberRange): Unit = {
+ seqNumRangesInCurrentBlock += range
+ }
+
+ /**
+ * Finalize the ranges added to the block that was active
+ * and prepare the ranges buffer for next block.
+ */
+ private def finalizeRangesForCurrentBlock(blockId: StreamBlockId): Unit
= {
+ blockIdToSeqNumRanges(blockId) = seqNumRangesInCurrentBlock.toArray
+ seqNumRangesInCurrentBlock.clear()
+ logDebug(s"Generated block $blockId has ${
blockIdToSeqNumRanges(blockId).mkString(", ")}")
+ }
+
+ private def storeBlockWithRanges(
--- End diff --
Good catch. How about making the `blockIdToSeqNumRanges` implement
`SynchronizedMap`?
The reason I dont want to add synchronized before called onPushBlock is
because its a bad idea to synchronized long-running block of code
(Receiver.store() can take a long time to return from storing the block). The
storing of the block really does not need be done under a lock, only the
accessing and updating of `blockIdToSeqNumRanges` needs to be done in a
synchronized manner.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
