I have a large cluster running on AWS. I keep having troubles with the disk space on my minions w.r.t. the "built-in" pods es-logging and influx-db. (I already hacked the cluster scripts to add use bigger ephemeral storage on instance creation.)

For influx-db, I just added an EBS volume and redeployed the controller. This seems to work for now.

For the elasticsearch-logging controller, I have a bit of an issue. Per default, there are two instances running. How can I specify the EBS volumes in the controller yaml file so that I don't run into the issue of trying to attach a single EBS volume to two pods.

Does K8S provide a solution to this? How is controller scaling and attached storage handled in general?

I hope I could make that somewhat clear ;).