Iv'e recently enabled compaction on Hive and I'm seeing a job appear regularly in the default queue. I'd like to portion off a slither of the cluster to this function by creating a new queue for it but how do I tell it what queue to use to do this rather than it going to default?
Compaction will not honour hive.compactor.job.queue. Basically compactions works differently for fully acid tables and insert only tables.
For fully acid tables, when you perform a manual/auto compaction, there are two jobs spawned, one MR which is responsible for compaction which will honour compaction queue and another tez job which is responsible for stats analysis and is a tez job submitted to default queue.
For inserts only tables, when you perform a manual/auto compaction, there is tez job spawned which is submitted to default queue.
There is a jira raised raised which is being worked on it. Bug details for your reference.