is it possible to let the elastic-operator deploy multiple data nods/pods at once?
We have 24 data nodes being evenly distributed on 3 kubernetes worker. We’d like to deploy multiple at once because otherwise it takes like 70 minutes currently.
Reducing the amount of data nodes isn’t possible, at least in the short term.
It’s not the PDB, so I assume it’s the operator itself that steers that?
Hello, I'm not sure to understand your question because only “master" nodes are added/removed one by one. Could you provide more details about what you want to achieve?
So when we do config changes / re-deploy or update the operator, it will recreate all pods and that one by one. We have 24 data pods/nodes and that’s taking quite some time.
We’d like to let it deploy multiple ones at once. Whoever controls that.
You can see the age on the right side where you see that it starts with the last one first and goes up to data-0 which is still todo. So anything older ~1h is still to be deployed/recreated.
Predicates that prevent Pods from being recreated are logged by the operator during the upgrade. This should help explain why only one Pod can be deleted at any given time.
{"log.level":"info","@timestamp":"2025-08-28T08:45:56.225Z","log.logger":"elasticsearch-controller","message":"Cannot restart some nodes for upgrade at this time","service.version":"3.1.0+9fc9bb06","service.type":"eck","ecs.version":"1.4.0","iteration":"492","namespace":"siem","es_name":"siem","namespace":"siem","es_name":"siem","failed_predicates":{"if_yellow_only_restart_upgrading_nodes_with_unassigned_replicas":["siem-es-data-0","siem-es-data-1","siem-es-data-10","siem-es-data-2","siem-es-data-3","siem-es-data-4","siem-es-data-5","siem-es-data-6","siem-es-data-7","siem-es-data-8","siem-es-data-9","siem-es-masters-0"]}}
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.