Closed Panlichen closed 3 years ago
for the first error: the CNI config is invalid according to the latest CNI specification. as the error message explains, it is mandatory to add "name" attribute to a CNI config file you can add it with literally any value, cause nobody is reading it or using it for anything (yet it is mandatory)
for the second error: you don't have a default network configured in your cluster, and you haven't specified any network connections for your Pod so DANM has no idea what do you want to do with this guy :) please read https://github.com/nokia/danm/blob/master/user-guide.md#defining-default-networks
Thanks for your reply.
After a lot of work, I am now able to use SRIOV in Kubernetes with sriov-operator, and I have to push forward my project so I may not try to deal with danm recently. However, the philosophy behind danm is quite interesting, I am sure I'll be back.
Is this a BUG REPORT or FEATURE REQUEST?:
bug
What happened:
I use danm-installer job to deploy danm, and succeed, as shown:
I deploy a simple danm network:
and simple pod:
the net seems good:
Stuff in /etc/cni/net.d is:
but the pod gets stuck:
It seems that the bootstrap Network is not properly configured and the pod cannot run.
After creating the cluster with kubeadm init/join, I deploy flannel CNI as the official says:
and it creates 10-flannel.conflist in /etc/cni/net.d, which seems cannot be used by danm directly.
So in danm-installer-config.ymal, I use
to let danm create flannel.conf in /etc/cni/net.d.
Besides, when I deploy a normal pod like
it also stucks, and the error message is more clear:
Could you please help me to figure out what the error is actually about and how I can fix it?
What you expected to happen:
Run smoothly.
How to reproduce it:
Anything else we need to know?:
Environment:
DANM version (use
danm -version
):Kubernetes version (use
kubectl version
):DANM configuration (K8s manifests, kubeconfig files, CNI config file):
OS (e.g. from /etc/os-release):
Kernel (e.g.
uname -a
):Others: