RQG: Secondary indexes creating using defer_build=True take really long to get removed after deleting the bucket

Hi John

I know we have discussed this before also.

But I am trying to get all rqg jobs run in parallel for 1000 queries.

There are few jobs which create more than 20 secondary indexes using defer_build=True.

My wait time for rerunning rqg jobs has increased a lot owing to this improvement I am filing.

When I delete a particular bucket from a 1 node cluster, it taken more than 15 mins for the secondary indexes which were in READY state to get removed.

Hence I am not able to run any other rqg job on same vm since indexer complains that its in busy state.

Also when we query : select * from system:indexes; , it returns empty results.

Is there a way to improve these timings?