when add --vnet-subnet-id paramter, the karpanter not

[ottodeng]

Version

az aks create --name pix-nap-sg03 -l southeastasia
--resource-group pix-nap
--node-provisioning-mode Auto
--network-plugin-mode overlay
--network-dataplane cilium
--max-pods 110 --node-count 3
--service-cidr 10.97.96.0/20 --dns-service-ip 10.97.96.10
--network-plugin azure --enable-managed-identity
--node-osdisk-size 512 --node-vm-size Standard_D2s_v4
--vnet-subnet-id /subscriptions/2a5e7fa2-d528-4c66-9e6a-3a3a0f290e98/resourcegroups/pix-nap/providers/Microsoft.Network/virtualNetworks/aks-nap-vnet-junfeng/subnets/aks
--ssh-key-value ~/.ssh/id_rsa.pub
--tier premium --k8s-support-plan AKSLongTermSupport --kubernetes-version 1.27 --yes

Expected Behavior

nodeclaim will created success.

Actual Behavior

no any nodeclaim create.

(base) [email protected]:/Users/ottodeng/aks $ kubectl get nodeclaims.karpenter.sh
No resources found
(base) [email protected]:/Users/ottodeng/aks $ kubectl get events -A --field-selector source=karpenter
No resources found
(base) [email protected]:/Users/ottodeng/aks $

(base) [email protected]:/Users/ottodeng/aks $ kubectl get pod
NAME READY STATUS RESTARTS AGE
pix-88d6475c8-kt8gb 0/1 Pending 0 7m54s
(base) [email protected]:/Users/ottodeng/aks $ kubectl describe pod pix-88d6475c8-kt8gb
Name: pix-88d6475c8-kt8gb
Namespace: default
Priority: 0
Service Account: default
Node:
Labels: app=samples-tf-mnist-demo
pod-template-hash=88d6475c8
Annotations:
Status: Pending
IP:
IPs:
Controlled By: ReplicaSet/pix-88d6475c8
Containers:
samples-tf-mnist-demo:
Image: mcr.microsoft.com/azuredocs/samples-tf-mnist-demo:gpu
Port:
Host Port:
Args:
--max_steps
50000
Limits:
nvidia.com/gpu: 1
Requests:
nvidia.com/gpu: 1
Environment:
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-f7tj8 (ro)
Conditions:
Type Status
PodScheduled False
Volumes:
kube-api-access-f7tj8:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
ConfigMapOptional:
DownwardAPI: true
QoS Class: BestEffort
Node-Selectors:
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
nvidia.com/gpu:NoSchedule op=Exists
pool-type=t4:NoSchedule
sku=gpu:NoSchedule
Events:
Type Reason Age From Message

---

Warning FailedScheduling 2m37s (x4 over 8m7s) default-scheduler 0/3 nodes are available: 3 Insufficient nvidia.com/gpu. preemption: 0/3 nodes are available: 3 No preemption victims found for incoming pod..
(base) [email protected]:/Users/ottodeng/aks $

Steps to Reproduce the Problem

az aks create --name pix-nap-sg03 -l southeastasia
--resource-group pix-nap
--node-provisioning-mode Auto
--network-plugin-mode overlay
--network-dataplane cilium
--max-pods 110 --node-count 3
--service-cidr 10.97.96.0/20 --dns-service-ip 10.97.96.10
--network-plugin azure --enable-managed-identity
--node-osdisk-size 512 --node-vm-size Standard_D2s_v4
--vnet-subnet-id /subscriptions/2a5e7fa2-d528-4c66-9e6a-3a3a0f290e98/resourcegroups/pix-nap/providers/Microsoft.Network/virtualNetworks/aks-nap-vnet-junfeng/subnets/aks
--ssh-key-value ~/.ssh/id_rsa.pub
--tier premium --k8s-support-plan AKSLongTermSupport --kubernetes-version 1.27 --yes

Resource Specs and Logs

no any karpenter log.

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

[Bryce-Soghigian]

#199

[Bryce-Soghigian]

We are in the process of preparing an image for "--vnet-subnet-id" support as we speak.

[Bryce-Soghigian]

Duplicate of #199, we are in the process of rolling out this support right now for node auto provisioning. Self-hosted karpenter has support for this change already. Thank you for your patience!

[Bryce-Soghigian]

#238 is the PR that added support.