Is this a BUG REPORT or FEATURE REQUEST? (choose one): BUG REPORT
Minikube version (use minikube version
): v0.13.1
Environment:
cat ~/.minikube/machines/minikube/config.json | grep DriverName
): virtualbox and xhyvedocker -v
): Docker version 1.12.5, build 7392c3bWhat happened:
I was running minikube using virtualbox and wanted to change to use native xhyve. Even after specifying --vm-driver xhyve option, it still started existing virtualbox VM. Debug logging showed it is defaulted to old driver.
I had to delete machine explicitly and then deploy again using xhyve
What you expected to happen:
It should ignore existing deployment and use provided option.
OR give a warning saying defaulting to old driver as machine exists.
How to reproduce it (as minimally and precisely as possible):
Start minikube using no vm-driver option and then without deleting machine try deploying using xhyve.
If theres a running machine, we can't really switch the driver on the fly. We should definitely put a warning if the machine is running with different options than what was passed in.
I don't think we're going to implement this.
I agree with r2d4, it's rather confusing that the only way to find out which driver was used is to pay attention to top
or Activity Monitor to see which processes are eating cycles and memory. It would be helpful to issue a warning, or even something as simple as printing which driver is used instead of the vague Starting VM...
message.