Troubleshooting
StackState for Kubernetes troubleshooting Self-hosted
Here is a quick guide for troubleshooting the startup of StackState:
- 1.Check that the install completed successfully and the release is listed:helm list --namespace stackstate
- 2.Check that all pods in the StackState namespace are running:kubectl get podsIn a first deployment it can be that containers in several pods restart a few times, because they are waiting for other pods to start up and be in the
readystate. This can be delayed due to scheduling and docker image pulling delays.Pods that are inpendingstate are usually an indication of a problem:- The pod is unschedulable due to lack of resources in the cluster. If a cluster auto-scaler is active it will often be able to resolve this automatically, otherwise manual intervention is needed to add more nodes to the cluster
- The pod is unschedulable, there are nodes it would fit on, but those nodes have
taintsthat the pod doesn not tolerate. To solve this more nodes can be added that don't have the taints, but StackState can also be configured to tolerate certain taints and run on the tainted nodes. - The pod is waiting for the Persistent Volumes (PVs) to be mounted. A cause can be that the StackState Helm chart doesn't specify a
storageClassNamebut relies on the cluster having a default storage class. When there is no default for the cluster it's required to specify a storage class via the Helm values of StackState.
For pods with stateImagePullBackOffalso check the exact error message, common causes are:- An incorrect username/password used to pull the images
- Connecting to the docker registry failed, this can be due to authentication issues or connectivity issues (firewalls, air-gapped installations)
- A typo in the overriden docker image registry URL
To find out a more detailed cause for thePending,ImagePullBackOfforCrashLoopBackOffstates use this command:kubectl describe pod<pod-name>The output contains aneventsection at the end which usually contains the problem. It also has aStatesection for each container that has more details for termination of the container. - 3.
- 4.
Last modified 1mo ago