We have a cluster with the following configurations
- 12 nodes
- Elasticsearch 5.x
- 24 TB data
I need to upgrade this cluster to Elasticsearch 7.x
So we have come up with a plan to provision a new cluster with exact same hardware configurations as that of the original.
We need to re-index the data from the 5.x to 7.x cluster.
We are considering two options for that
- using a single powerfull logstash machine
- using the re-index API from the 7.x cluster
Which one is a better choice?
Does the re-index API makes use of all nodes (all nodes are master eligible, ingest and data nodes by default) and fare better than a single logstash machine?