Ambiguity in updating existing data with arbitrary segments from druid.coordinator.merge.on


I’m using druid.coordinator.merge.on which merges segments to arbitrary long segments, time wise.

When updating these segments I’m not quite sure how to setup hadoop indexing job.

IngestionSpec.IoConfig.IngestionUpdateSpec takes :

  • intervals - does it serve simply as time restriction, so that I can put single interval that spans all segments that I want to reindex ?

  • segments - if not specified, get resolved using /druid/coordinator/v1/metadata/datasources/segments

IngestionSpec.DataSchema.ArbitraryGranularitySpec takes :

  • intervals - what about this one, does it also serve as a time restriction ?

As you can see, I’m not quite sure what to put into Intervals to reindex all segments that I have in data-source.

and use it for intervals in both IngestionSpec.IoConfig.IngestionUpdateSpec and IngestionSpec.DataSchema.ArbitraryGranularitySpec

I guess that if I want to preserve the intervals of segments, I should get /druid/coordinator/v1/metadata/datasources/segments myself