Troubleshooting

This guide helps you troubleshoot common issues with runwasi shims.

Container Fails to Start

If your wasm container fails to start, you might see events like these:

Events:
  Type     Reason                  Age               From               Message
  ----     ------                  ----              ----               -------
  Normal   Scheduled               18s               default-scheduler  Successfully assigned default/wasi-demo-75d6bb666c-479km to minikube
  Warning  FailedCreatePodSandBox  4s (x2 over 18s)  kubelet            Failed to create pod sandbox: rpc error: code = Unknown desc = failed to get sandbox runtime: no runtime for "wasmtime" is configured

Investigation Steps:

Check runtime class configuration: Ensure your RuntimeClass is correctly configured.
```
kubectl get runtimeclass

# Example output:
 NAME                  HANDLER               AGE
 wasm                  wasm                  8d
 wasmedge              wasmedge              8d
 wasmer                wasmer                8d
 wasmtime              wasmtime              8d
```
RuntimeClass in Kubernetes defines a handler name (e.g., wasmtime) that maps directly to a CRI section in containerd’s config.toml ([plugins."io.containerd.grpc.v1.cri".containerd.runtimes.wasmtime]), which then points to the shim binary (containerd-shim-wasmtime-v1) that executes your Wasm workload. See more details in Runtime classes from containerd

Check containerd configuration: Verify that the containerd configuration has a runtime configured for your Wasm shim (e.g., wasmtime).

cat /etc/containerd/config.toml

# Look for a section like:
[plugins."io.containerd.grpc.v1.cri".containerd.runtimes.wasmtime]
   runtime_type = "io.containerd.wasmtime.v1"

Make sure the containerd shim binary is in the PATH.

Log-Based Troubleshooting

Examining logs is one of the most effective ways to troubleshoot issues in runwasi shims.

Finding containerd Logs

Logs can be in different locations depending on your system:

Kubernetes Distributions

k3s: /var/lib/rancher/k3s/agent/containerd/containerd.log
k8s (kubeadm): /var/log/containerd/containerd.log

Using journalctl (systemd-based systems)

# View all containerd logs
journalctl -u containerd

# Follow containerd logs
journalctl -u containerd -f

Kubernetes Pod Logs

In addition to containerd logs, Kubernetes pod logs show the logs from your containers.

kubectl logs <pod-name>

# Or stream logs in real-time
kubectl logs -f <pod-name>

Using Structured Logging with Container and Pod IDs

runwasi shims include structured logging that automatically adds container ID and pod ID to log messages, making it easier to filter logs.

Finding Container ID and Pod ID:

# Get container IDs for a pod
kubectl get pod <pod-name> -o jsonpath='{.status.containerStatuses[*].containerID}'
# Output: containerd://<container-id>

# Get pod ID for a specific pod
crictl pods --name <pod-name> -q

Note: If you are using k3s, you can use the following command to get the pod ID:

k3s crictl pods --name <pod-name> -q

Filtering Logs by Container ID (k3s)

grep 'instance="<container-id>"' /var/lib/rancher/k3s/agent/containerd/containerd.log

# Example output:
time="2025-03-05T21:49:01.527630395Z" level=info instance="820f78385e9c29cbd6a0b6767619286ffd7a3384959ce909063a22041d17c718" pod="22a0daacfeabe74d165552794ec9615c67545566bec282d3fe0b5cc910e9cdb5" msg="setting up wasi"

Filtering Logs by Pod ID (k3s)

grep 'pod="22a0daacfeabe74d"' /var/lib/rancher/k3s/agent/containerd/containerd.log

Runwasi Developer Documentation