There are a couple of ways we might want to do this:
1) just take args for all sizes (master, node, infra)
2) calculate disk size based on environment and whether or not we're provisioning the smoke test stuff.
the assumption here is that if you are doing the smoke test stuff you're probably doing lots of labs which might require more disk
I had bumped the infra disk to 60g of io1 back when we were having bad performance, but the Docker backport and registry fixes may have made all of that unnecessary.
There are a couple of ways we might want to do this:
1) just take args for all sizes (master, node, infra) 2) calculate disk size based on environment and whether or not we're provisioning the smoke test stuff. the assumption here is that if you are doing the smoke test stuff you're probably doing lots of labs which might require more disk
I had bumped the infra disk to 60g of io1 back when we were having bad performance, but the Docker backport and registry fixes may have made all of that unnecessary.