WIP

This article is a work in progress.

Target Audience

This page is primarily intended for users on Private Cloud plans. If you are using Metaplay Cloud with a Pre-launch/Production plan, information from this page may not be directly relevant to your needs.

Dependencies

• kubectl (v1.18 or later)
• Helm (v3.9 or later)
• Prior knowledge of Debug Cluster is highly recommended.
• A valid kubeconfig.yaml

Common Faults and Quick Fixes

Nodes

• Nodes are not in Ready states: This usually occurs when some pods on the node have consumed too much memory, causing the kubelet DaemonSet to die and no longer report back to the EKS control plane. In such cases, you need to manually remove the node in question from the cluster and wait for a new node to be provisioned as a replacement.

Pods

• Pod in Pending state forever: You can describe the pod and check its Events to find out the exact issue that prevents a pod from being properly assigned. A common reason is that the Pod is requesting too many resources, exceeding the quota.

Attaching Debug Container to Game Server Pod

Metaplay has provided you with a debug container metaplay/diagnostics which can be used for advanced debugging against a running game server pod:

kubectl debug <pod name> -it --profile=general --image metaplay/diagnostics:latest --target shard-server -n <namespace>

LiveOps Dashboard

• 401 Error: Please ensure that you have the correct role from your authentication provider.
• 404 Error: This could happen due to DNS propagation delay after a freshly made deployment. If the issue occurs for more than 10 minutes, please check the Kubernetes Ingress/Service and Route53 DNS records for further investigation.
• 503 Error: The LiveOps Dashboard is shipped with the game server image; please check the game server pod status.