I have deployed and experimented with OpenShift 3.11 in the past few weeks, and I am now moving to a new deployment from scratch past this period.
Still, I have a few questions about storage I would like to share if you can give me some opinions.
The deployment is being made in VPS hosted at a provider we trust. We have very good plans that make us want to stick to it. This offering has a catch: we only have access to some pre-determined VPS configs, and we can’t change these configs or we’ll lose the price benefit that matters in the first place.
So here the thing is they all come with a great amount of storage, varying between 400GB and 2TB per VPS.
Since I am implementing Master -> Infra -> compute, I have a surplus of local storage, being that this storage is always one drive only. It can be partitioned, but it can never be a separate block device.
Now, what I would like to know is, given the above, what would be the better option for persistent storage, and what engine for automatic provisioning to use?
Should it be NFS, FlexVolume or other?
Where will the storage make sense and where will it not? Does it make sense to have the storage on the Master, or would it make more sense in some Infra or Compute node? What performance issues may I expect, and are they generally neglectable or noticeable?
Thank you all!