Currently we are using Elastic Search 2.3.1 and trying to issue XDELETE command on the indices. I understand it's pretty old version & has attained EOL but I assume the underlying working mechanism of XDELETE wouldn't change!
Background information -
- ES is deployed on 6 cluster nodes
- We use NAS for ES data storage
The command we are issuing is curl -XDELETE http://localhost:9200/indexName/ - This doesn't delete from all the data nodes. Only gets deleted from few set of nodes & the issue happens randomly. Randomly in the sense some times XDELETE deletes from all the nodes of the cluster but sometimes it doesn't delete from all the nodes & the data stays in few of the nodes.
I have the below queries
- I did check the Elastic Search logs in the specific nodes where the underlying data is not getting deleted but didn't find any messages related to delete. Let me know if any other logs to check here?
- Also want to know the working of XDELETE on Linux / OS level - What are commands underlying does it uses? Does it issue any SSH before deleting? Will it issue any rm -rf Linux command in turn to delete to folders from the ES node directories?