I wanted to remove a Shard but after running

db.adminCommand({ removeShard: 'shard_07'})

 draining started but never finished. 

After some time I found the reason. For some collections I did a sh.disableBalancing('<namespace>') which caused the draining to hang.

Of course, this behavior is not a bug, however the user does not get any indication why shard removing never finishes.

It should be reflected in the output of removeShard command, maybe like this:

{
  msg: 'draining ongoing',
  state: 'ongoing',
  remaining: { chunks: Long('28'), dbs: Long('0'), jumboChunks: Long('0') },
  note: 'you need to drop or movePrimary these databases',
  dbsToMove: [],
  balancer_note: 'you need to enableBalancing on these collections',
  lockedCollections: ['data.sessions.20251112']
  ok: 1
}
 

At least, it should be mentioned in the documentation Remove Shards from a Sharded Cluster. Maybe like this:

Steps

1. Ensure the balancer is enabled
   ...
2. Ensure balancing is enabled on collections
   Run command

   db.getSiblingDB("config").collections.find({noBalance: true})

   If output is not empty, then run sh.enableBalancing('<namespace>') on these collections

1.  
   Migrate sharded collection data with the balancer
   ...