I am ingesting data in Real-time node via Kafka Firehose. After handing off the segment to historical node ( verified via logs and coordinator console), realtime node doesn’t delete some segments from the local tmp folder. Some segments are deleted after a lot of time ( after 9-10 hours).
What might be the cause for the delay?
Is there any way to force the deletion of such segments ?
How are tasks like handing-off and deletion of segments scheduled?