APISIX Helm Deployment on Minikube fails

[tomas-rojo]

I’m encountering persistent issues deploying the APISIX Helm chart (version apisix-2.10.0, app version 3.11.0) on a Minikube cluster. Despite resolving an initial storage class mismatch, the deployment remains unhealthy: etcd pods run but aren’t ready (0/1), the ingress controller hangs in Init:0/1, and the dashboard crashes repeatedly. I suspect compatibility or configuration issues with Minikube or the chart’s defaults. Here’s a detailed breakdown of my setup and troubleshooting steps.

Environment

• Chart Version: apisix-2.10.0
• App Version: 3.11.0
• Minikube: Started with minikube start -n=2 --driver=docker (2 nodes, Docker driver)
• Kubernetes Client Version: v1.30.2
• Kubernetes Server Version: v1.31.0
• Helm Command: helm install apisix apisix/apisix --create-namespace -n apisix -f apisix-values.yml

Helm Values (apisix-values.yml)

dashboard:
  enabled: true
service:
  type: NodePort
ingress-controller:
  enabled: true
  config:
    apisix:
      serviceName: release-name-apisix-admin
      adminAPIversion: v3
    kubernetes:
      enableGatewayAPI: true

Initial Pod Status

kubectl get po -n apisix

NAME                                         READY   STATUS             RESTARTS      AGE
pod/apisix-6699fffc65-78t6v                      0/1     Init:0/1           0               20m
pod/apisix-dashboard-5986b9dc66-hh6dh            0/1     CrashLoopBackOff   8 (4m1s ago)    20m
pod/apisix-etcd-0                                0/1     CrashLoopBackOff   6 (2m26s ago)   20m
pod/apisix-etcd-1                                0/1     CrashLoopBackOff   6 (2m17s ago)   20m
pod/apisix-etcd-2                                0/1     Running            6 (2m48s ago)   20m
pod/apisix-ingress-controller-57866db6fc-v88ld   0/1     Init:0/1           0               20m

k describe po apisix-ingress-controller-57866db6fc-v88ld -n apisix

Events:
  Normal   Scheduled    3m54s  default-scheduler  Successfully assigned ... to minikube-m02
  Warning  FailedMount  3m53s  kubelet            MountVolume.SetUp failed for volume "kube-api-access-48b9d" : failed to sync configmap cache: timed out waiting for the condition
  Normal   Pulled       3m52s  kubelet            Container image "busybox:1.28" already present on machine
  Normal   Created      3m52s  kubelet            Created container wait-apisix-admin
  Normal   Started      3m51s  kubelet

Current State

• etcd Pods: Running but not ready (0/1) with restarts.
• Ingress Controller: Stuck in Init:0/1
• The FailedMount suggests a Minikube issue, but the init container (wait-apisix-admin) started, implying it’s waiting for APISIX admin service.
• Dashboard: CrashLoopBackOff. Likely waiting for APISIX, which depends on etcd.
• APISIX: Stuck in Init:0/1, likely waiting for etcd to become ready.

Could this be a bug in the Helm chart’s defaults or Minikube compatibility?
I’m happy to test suggestions or provide more output. Thanks for any insights!