After upgrade my elasticsearch node selectors are gone

Issue

• We decided after experiencing very bad performances with NFS (Not recommended for Lucene) to switch to local storage on dedicated nodes
• Due to the node label selector which should be dedicated per deployment configs so as to avoid having the pod moving from nodes to nodes.
• It is not so obvious in the documentation as the openshift_logging_es_nodeselector expects a single value and therefore that logic needs to be done post-deployment and not covered by the Ansible automation.