Github user vivekmittal commented on a diff in the pull request:
https://github.com/apache/storm/pull/2104#discussion_r115406125
--- Diff:
external/storm-kafka-client/src/main/java/org/apache/storm/kafka/spout/internal/OffsetManager.java
---
@@ -68,13 +74,34 @@ public OffsetAndMetadata findNextCommitOffset() {
KafkaSpoutMessageId nextCommitMsg = null; // this is a
convenience variable to make it faster to create OffsetAndMetadata
for (KafkaSpoutMessageId currAckedMsg : ackedMsgs) { //
complexity is that of a linear scan on a TreeMap
- if ((currOffset = currAckedMsg.offset()) == nextCommitOffset +
1) { // found the next offset to commit
+ currOffset = currAckedMsg.offset();
+ if (currOffset == nextCommitOffset + 1) { // found
the next offset to commit
found = true;
nextCommitMsg = currAckedMsg;
nextCommitOffset = currOffset;
- } else if (currAckedMsg.offset() > nextCommitOffset + 1) {
// offset found is not continuous to the offsets listed to go in the next
commit, so stop search
- LOG.debug("topic-partition [{}] has non-continuous offset
[{}]. It will be processed in a subsequent batch.", tp, currOffset);
- break;
+ } else if (currOffset > nextCommitOffset + 1) {
+ if (emittedOffsets.contains(nextCommitOffset + 1)) {
+ LOG.debug("topic-partition [{}] has non-continuous
offset [{}]. It will be processed in a subsequent batch.", tp, currOffset);
+ break;
+ } else {
+ /*
+ This case will arise in case of non contiguous
offset being processed.
+ So, if the queue doesn't contain offset =
committedOffset + 1 (possible
+ if the queue is compacted or deleted), the
consumer should jump to
+ the next logical point in the queue. Next logical
offset should be the
+ first element after committedOffset in the
ascending ordered emitted set.
+ */
+ LOG.debug("Processed non contiguous offset, the
previously committed offset has been deleted from the topic. Committed: [{}],
Processed: [{}]", committedOffset, currOffset);
--- End diff --
Yes the offset is there because it was part of the topic, but there is no
guarantee that it will always be there. The offsets might be wiped off owing to
compaction/deletion process in a kafka topic. This case will only arrive when
the previously (committedOffset + 1) is no longer present in the topic. I will
rephrase the message to explicitly specify (committedOffset+1) is no longer
part of the topic.
---
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.
---
