While using rethinkdb import to a 9 node cluster on a table with 150m rows during a reconfigure the web interface became unusable, presumably due to very long calls reading the rethinkdb.table_status table.

r.db('rethinkdb').table('table_status').count() took 2min 5.61s., but did return the correct result of 16 tables.

The reconfigure was going from 9 shards/2 replicas -> 8 shards/2 replicas. The interface became unusable 10-15 minutes after the import was killed. 1 replica is in the 'backfilling' state, all others are in the 'ready' state.

"status": {
"all_replicas_ready": false ,
"ready_for_outdated_reads": true ,
"ready_for_reads": true ,
"ready_for_writes": true
}