Speeding up Elasticsearch snapshot restore in ES 5.2

We are using a 5.2 ES cluster and trying to restore a lot of data from AWS s3. The process is going slower than we would like so we have increased the following settings:

cluster.routing.allocation.node_concurrent_recoveries (from I believe the default of 2 to 4)
indices.recovery.max_bytes_per_sec (from 40mb to 120mb)

Besides changing these two settings is there anything else we can tweak in order to increase our restore speed? (Did we even change the correct settings in the first place?)

As this is a restore from a snapshot, you should also increase the repository setting max_restore_bytes_per_sec (which is set by default to 40mb), see https://www.elastic.co/guide/en/elasticsearch/reference/current/modules-snapshots.html

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.