The Mac xhyve driver appears to ignore --memory and --cpus switches:
->minikube start --vm-driver=xhyve --container-runtime=docker --show-libmachine-logs --v=10 --alsologtostderr --cpus 4 --memory 6120
[snip]
(minikube) DBG | [xhyve -A -U D6603CF9-B0C5-4B12-81A1-3AC2827DA143 -c 2 -m 2048M -l com1,autopty ....]
It doesn't look like the command-line args are getting passed through to xhyve correctly; I would expect to see -c 4 -m 6120M
Minikube version:
minikube version: v0.14.0
Environment:
This is more of an issue with user experience. If the minikube VM is already started, memory and cpus flags won't have any effect as hyperkit cant change some options without a recreate. Right now it silently ignores the flags if minikube is already started.
We've debated whether to have some help text and continue to ignore the flags or to split minikube start into minikube create and minikube start
Confirmed. Using delete before trying to start again shows the correct settings. Thanks for the clarification @r2d4
FWIW a message like this would be quite clear:
WARNING: Memory and cpu options are ignored when resuming an already-created minikube instance.
+1 @dbanttari , having a warning would definitely save people time when trying to allocate more resources to the VM. Ignoring flags is not a great way to go. This bit me on ubuntu 14.04.
I got bitten by this today - was warning text idea shelved?
I believe the same applies to vbox. A warning message would be very much welcome.
Most helpful comment
Confirmed. Using
deletebefore trying tostartagain shows the correct settings. Thanks for the clarification @r2d4FWIW a message like this would be quite clear: