Exception on range partitioning during compaction

[sgaragan]

Please provide a detailed title (e.g. "Broker crashes when using TopN query with Bound filter" instead of just "Broker crashes").

Affected Version

Druid 30.0.1

Description

I noticed an exception in the task logs for a coordinator issued compaction for a table with a lot of large rows, mostly from one column with a JSON encoded string. The processing continued as this seems to be a problem with the stats collection as shown by the exception

java.lang.NullPointerException: Cannot invoke "java.util.Map.get(Object)" because "ingestionStatsAndErrors" is null at org.apache.druid.indexing.common.task.batch.parallel.ParallelIndexSupervisorTask.getRowStatsAndUnparseableEventsForRunningTasks(ParallelIndexSupervisorTask.java:1651) ~[druid-indexing-service-28.0.1.jar:28.0.1] at org.apache.druid.indexing.common.task.batch.parallel.ParallelIndexSupervisorTask.doGetRowStatsAndUnparseableEventsParallelMultiPhase(ParallelIndexSupervisorTask.java:1623) ~[druid-indexing-service-28.0.1.jar:28.0.1] at org.apache.druid.indexing.common.task.batch.parallel.ParallelIndexSupervisorTask.doGetRowStatsAndUnparseableEvents(ParallelIndexSupervisorTask.java:1724) ~[druid-indexing-service-28.0.1.jar:28.0.1] at org.apache.druid.indexing.common.task.batch.parallel.ParallelIndexSupervisorTask.doGetLiveReports(ParallelIndexSupervisorTask.java:1760) ~[druid-indexing-service-28.0.1.jar:28.0.1] at org.apache.druid.indexing.common.task.batch.parallel.ParallelIndexSupervisorTask.getLiveReports(ParallelIndexSupervisorTask.java:1795) ~[druid-indexing-service-28.0.1.jar:28.0.1]

The full exception is attached to this issue

This cluster is running on OpenShift using the Druid Operator with 1 broker 1 coordinator/overlord 1 router 7 historicals 4 middle managers

task-compaction-log.log