I have an ES node, which has temporarily attached EBS volumes belonging to our former cluster which went down due to operator error. Attaching these EBS volumes, and adding them to path.data has allowed fairly easy reallocation to the other nodes of the cluster, so that I can decommission these older, slower (st1) EBS volumes. The issue that I have now is that there are about 700 closed indices (all log data from months ago).
I've tried opening a small set of closed indices at a time, and letting it reallocate. This caused the cluster to be "red" for too long. That at doing this process x700, x100, etc. would not be fun. Maybe someone has a better suggestion how how to handle this?
Perhaps I should manually copy these indices deep directories in the filesystem to the newer node's SSD volumes and restart the ES node? Tell me if this is a terrible idea or not!