Open mullerlele opened 4 years ago
Thanks for reporting this issue @mullerlele. I was not able to reproduce this issue on AKS with a 1.14.7 cluster. The image mcr.microsoft.com/k8s/flexvolume/keyvault-flexvolume:v0.0.15
was pulled successfully on each node. Can you pls try again?
Thank you @ritazh for your prompt reply. It still behaves the same - did you try it on a cluster with multiple vmss and aci as a virtual node? (the aci virtual node should not be of any concern in this case, because it is labeled not to have pod scheduled there, but I am only mentioning it here to provide a more detailed context)
Hi @mullerlele I tried to repro this but could not. Could you pls provide more detailed repro steps. I followed the steps in this doc to create a new AKS cluster. Then ran the following cmds:
kubectl get no
NAME STATUS ROLES AGE VERSION
aks-nodepool1-36484348-vmss000000 Ready agent 65m v1.14.7
aks-nodepool1-36484348-vmss000001 Ready agent 65m v1.14.7
aks-nodepool1-36484348-vmss000002 Ready agent 65m v1.14.7
virtual-node-aci-linux Ready agent 6m8s v1.13.1-vk-v0.9.0-1-g7b92d1ee-dev
kubectl apply -f https://raw.githubusercontent.com/Azure/kubernetes-keyvault-flexvol/master/deployment/kv-flexvol-installer.yaml
namespace/kv created
daemonset.extensions/keyvault-flexvolume created
kubectl get pod --all-namespaces -o wide
NAMESPACE NAME READY STATUS RESTARTS AGE IP NODE NOMINATED NODE READINESS GATES
kv keyvault-flexvolume-7bsg7 1/1 Running 0 33s 10.240.0.69 aks-nodepool1-36484348-vmss000002 <none> <none>
kv keyvault-flexvolume-sv96c 1/1 Running 0 33s 10.240.0.40 aks-nodepool1-36484348-vmss000001 <none> <none>
kv keyvault-flexvolume-vl2j9 1/1 Running 0 33s 10.240.0.23 aks-nodepool1-36484348-vmss000000 <none> <none>
...
Describe the bug installation failed on one of the nodes in the vmss, due to image pulling failures.
Steps To Reproduce 1) create a aks cluster with multiple vmss 2) provision more than 1 instance in one of the vmss (linux) 3) install flexvol based on the instructions 4) image pull back off (for the flexvol image) occurs on one of the node and hence the volume is not present on that node Expected behavior flexvol to be deployed as Daemonset with no problem
Key Vault FlexVolume version 0.0.15 Access mode: service principal or pod identity SP
Kubernetes version 1.14.7 Additional context