during my experiment:1.0.0-beta plugin is basic test that can work in two nodes cluster, and worker node GPU can be found correctly, but I think V0.6.0 Branch should be OK too. if you can can create cluster correctly ,omit post-install.sh procedure
next week ,I will recreate 3 nodes cluster and using V0.6 to repeat your test , please wait
Thanks so much for your patience! We are waiting for your update.
when I using yml gpu-test.yml you typed to create , it is failed with error:CrashLoopBackOff,
is this the complete yml you tested?
Yes, it is complete, and it runs the same as the following command. @Jeffli
kubectl run -i -t nvidia --image=jitteam/devicequery