Delete unassigned shards in Elasticsearch

I had an elasticsearch server, which ran in single node mode. When dataset reached 1TB, I added second node and relocated couple of shards with reroute api. Now second node has 2 of 5 shards, but first node still holds all 5 shards and space is not reclaimed. **_cat/shards?v** command shows: new_messages 3 p STARTED 974698739 256.6gb 5.188.130.61 el01 new_messages 3 r UNASSIGNED I've found some "solutions" like stop ES and delete files by hand, but I don't like them.