The snapshot restore should definitely be faster - but if you're set on reindexing you can rig that up pretty easily in the python client. See https://elasticsearch-py.readthedocs.io/en/master/helpers.html#reindex
from elasticsearch import Elasticsearch from elasticsearch.helpers import reindex es_source=Elasticsearch("source_system:9200") es_destination=Elasticsearch("destination_system:9200") reindex(es_source, "my_source_index", "my_destination_index", target_client=es_destination)
That's what I've done as I'm still stuck on 1.7.x - though the 2.3 reindex API looks handy.
Nik, when doing a local cluster->local cluster reindexing, is the reindex API in 2.3 faster/more efficient than the python client's approach, if the python client is run directly on the host cluster?