Unable to load CRI runtime service plugin dependency

[VxDlH]

Bug Report

talosctl -n kube-master-1 logs containerd

kube-master-1: {"id":"io.containerd.sandbox.store.v1.local","level":"info","msg":"loading plugin","time":"2024-09-24T20:39:48.040446278Z","type":"io.containerd.sandbox.store.v1"}
kube-master-1: {"id":"io.containerd.sandbox.controller.v1.podsandbox","level":"info","msg":"loading plugin","time":"2024-09-24T20:39:48.040457844Z","type":"io.containerd.sandbox.controller.v1"}
kube-master-1: {"error":"unable to load CRI runtime service plugin dependency: no plugins registered for io.containerd.cri.v1.runtime: plugin: not found","id":"io.containerd.sandbox.controller.v1.podsandbox","level":"warning","msg":"failed to load plugin","time":"2024-09-24T20:39:48.040854240Z","type":"io.containerd.sandbox.controller.v1"}
kube-master-1: {"id":"io.containerd.sandbox.controller.v1.shim","level":"info","msg":"loading plugin","time":"2024-09-24T20:39:48.040870396Z","type":"io.containerd.sandbox.controller.v1"}
kube-master-1: {"id":"io.containerd.grpc.v1.sandbox-controllers","level":"info","msg":"loading plugin","time":"2024-09-24T20:39:48.040929940Z","type":"io.containerd.grpc.v1"}
kube-master-1: {"error":"unable to load CRI runtime service plugin dependency: no plugins registered for io.containerd.cri.v1.runtime: plugin: not found","id":"io.containerd.grpc.v1.sandbox-controllers","level":"warning","msg":"failed to load plugin","time":"2024-09-24T20:39:48.040948417Z","type":"io.containerd.grpc.v1"}
kube-master-1: {"id":"io.containerd.grpc.v1.sandboxes","level":"info","msg":"loading plugin","time":"2024-09-24T20:39:48.040961737Z","type":"io.containerd.grpc.v1"}
kube-master-1: {"id":"io.containerd.grpc.v1.snapshots","level":"info","msg":"loading plugin","time":"2024-09-24T20:39:48.040975298Z","type":"io.containerd.grpc.v1"}

Description

How to fix the following errors :

kube-master-1: {"error":"unable to load CRI runtime service plugin dependency: no plugins registered for io.containerd.cri.v1.runtime: plugin: not found","id":"io.containerd.sandbox.controller.v1.podsandbox","level":"warning","msg":"failed to load plugin","time":"2024-09-24T20:39:48.040854240Z","type":"io.containerd.sandbox.controller.v1"}
kube-master-1: {"error":"unable to load CRI runtime service plugin dependency: no plugins registered for io.containerd.cri.v1.runtime: plugin: not found","id":"io.containerd.grpc.v1.sandbox-controllers","level":"warning","msg":"failed to load plugin","time":"2024-09-24T20:39:48.040948417Z","type":"io.containerd.grpc.v1"}

I have these errors on all nodes. I didn't declare this plugin, I think it's a default plugin, but why doesn't it load?

Logs

Environment

• Talos version: [talosctl version --nodes <problematic nodes>]

talosctl -n kube-master-1 version
Client:
        Tag:         v1.8.0
        SHA:         5cc935f7
        Built:       
        Go version:  go1.22.7
        OS/Arch:     darwin/arm64
Server:
        NODE:        kube-master-1
        Tag:         v1.8.0
        SHA:         5cc935f7
        Built:       
        Go version:  go1.22.7
        OS/Arch:     linux/amd64
        Enabled:     RBAC

• Kubernetes version: [kubectl version --short]

kubectl version
Client Version: v1.31.1
Kustomize Version: v5.4.2
Server Version: v1.31.1

• Platform: metal

[smira]

We can fix this warning, but it's totally fine - containerd service is a Talos system containerd instance, while the one you should be interested in is the cri service which runs containerd for Kubernetes, workloads, etc.

[VxDlH]

I totally agree with you. The only problem is that it's flagged as an error. Perhaps it would be better to classify it as a warning ? However, I do understand that it doesn't prevent proper operation.

At one point, I even thought the problem was with this configuration. Can you confirm whether this configuration is correct ?

machine:
  files:
    - content: |
        [plugins]
          [plugins."io.containerd.cri.v1.images".pinned_images]
            sandbox = "registry.k8s.io/pause:3.9"
      path: /etc/cri/conf.d/20-customization.part
      op: overwrite

[smira]

Your configuration has nothing to do with what you posted above, as you need to check talosctl logs cri.

Documentation.