Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
Description
When CheckpointUtils#computeOffsetRanges for consuming kafka messges.
Given
topic = "test",
fromOffsets(partition -> offset pair) = (0 -> 0), (1 -> 0), (2 -> 0), (3 -> 0), (4 -> 0),
toOffsets = (0, 100), (1, 1000), (2, 1000), (3, 1000), (4, 1000),
numEvents = 1001.
The output of CheckpointUtils#computesOffsetRanges is
OffsetRange(topic: 'test', partition: 0, range: [0 -> 100])
OffsetRange(topic: 'test', partition: 1, range: [0 -> 226])
OffsetRange(topic: 'test', partition: 2, range: [0 -> 226])
OffsetRange(topic: 'test', partition: 3, range: [0 -> 226])
OffsetRange(topic: 'test', partition: 4, range: [0 -> 226])
Total count is 1004(100 + 266 * 4), more than 1001, and thus consume more entries from kafka than specified 1001.
CC vinoth
Attachments
Issue Links
- links to