Closed GOVYANSONG closed 4 years ago
UPDATE:
I was trying to run the demo DIND. Everything has gone normally until the point for cirros-vm to be installed and created. Following is event logs for cirros-vm pod:
Events: Type Reason Age From Message
Warning Failed 55m (x247 over 119m) kubelet, kube-node-1 Error: ImagePullBackOff Warning Failed 5m1s (x25 over 119m) kubelet, kube-node-1 Failed to pull image "virtlet.cloud/cirros": rpc error: code = Unknown desc = "/run/virtlet.sock": rpc error: code = 2 desc = error downloading "https://github.com/mirantis/virtlet/releases/download/v0.9.3/cirros.img": Get https://github.com/mirantis/virtlet/releases/download/v0.9.3/cirros.img: dial tcp: i/o timeout Normal BackOff 9s (x471 over 119m) kubelet, kube-node-1 Back-off pulling image "virtlet.cloud/cirros"
It looks like that it is related to the original issue with /run/virtlet.sock.
I got a brand new ubuntu linux box, installed CRI proxy by following instructions given for Installation on kubeadm clusters for use with Virtlet, and deployed Virtlet DaemonSet as instructed. After removed taint from my single master node, virtlet pod was created successfully. Apologize for any confusion.
kubectl logs virtlet-8z5tn -n kube-system -c virtlet
UnixSocket Listener started Starting server on socket /run/virtlet.sock grpc: Server.Serve failed to create ServerTransport: connection error: desc = "transport: write unix /run/virtlet.sock->@: write: broken pipe" transport: http2Server.HandleStreams failed to read frame: read unix /run/virtlet.sock->@: read: connection reset by peer (<--note: this message repeats every 10 seconds)
Host: Ubuntu 18.04 K8S: 1.16.0 CNI: flannel DNS: CoreDNS
Docker: v19.03.2 Virtlet: vLatest or 1.5.1
Installation was undertaken by following instructions for systemd.
Could someone shed light on the root cause of this?