Upgrading the cluster using Elastic Operator


I am testing different upgrade options of Elasticsearch using Elastic Operator on Kubernetes following this documentation Upgrading the cluster. Example described there works fine:

  • Replace dedicated master and dedicated data nodes with nodes having both master and data roles: Replace the two existing NodeSets by a single one with a different name and the appropriate Elasticsearch configuration settings.

But I was wondering if the reverse use case is also supported (split combined master/data configuration to dedicated master and data nodes).

I tried to test it out, but I get the following error:

ReconciliationError elasticsearch/test Failed to apply spec change: adjust resources: Get "http://test-es-http.test-ns.svc:9200/_nodes/_all/no-metrics": dial tcp connect: connection refused

Any idea if this use case is supported or not?

I found the problem in my code. This case is supported as well, so this topic can be closed.