What is the defect and your suggestions on improvement:
When I tried to install fluid on a GKE cluster, I ran into several issues:
The default hostpath of kubelet kubeconfig path(/etc/kubernetes/kubelet.conf) is not right, for GKE the kubeconfig is in /var/lib/kubelet/kubeconfig
Error in csi-nodeplugin-fluid: unable to read certificate-authority /etc/srv/kubernetes/pki/ca-certificates.crt for default-cluster due to open /etc/srv/kubernetes/pki/ca-certificates.crt: no such file or directory, it is required to mount the /home/kubernetes/bin/gke-exec-auth-plugin and /etc/srv/kubernetes/pki/ca-certificates.crt into the csi-nodeplugin-fluid
Error in daemonset demo-fuse: Error creating: insufficient quota to match these scopes: [{PriorityClass In [system-node-critical system-cluster-critical]}], the helm value runtime.criticalFusePod should be set to false.
Provide a link to that doc page:
What is the defect and your suggestions on improvement:
When I tried to install fluid on a GKE cluster, I ran into several issues:
/etc/kubernetes/kubelet.conf
) is not right, for GKE the kubeconfig is in/var/lib/kubelet/kubeconfig
csi-nodeplugin-fluid
:unable to read certificate-authority /etc/srv/kubernetes/pki/ca-certificates.crt for default-cluster due to open /etc/srv/kubernetes/pki/ca-certificates.crt: no such file or directory
, it is required to mount the/home/kubernetes/bin/gke-exec-auth-plugin
and/etc/srv/kubernetes/pki/ca-certificates.crt
into thecsi-nodeplugin-fluid
Error creating: insufficient quota to match these scopes: [{PriorityClass In [system-node-critical system-cluster-critical]}]
, the helm valueruntime.criticalFusePod
should be set to false.