If a Helm deployment's status is failed, what can I do to determine what made it fail?
Asked
Active
Viewed 1.6k times
8
Nomnom
- 4,193
- 7
- 27
- 37
-
2What are the things you have tried already to determine this? – rock'n rolla Mar 31 '21 at 07:03
-
@rock'nrolla the nature of this question is general, I'm wondering if there's a canonical first few steps to be taken in order to determine deployment failures. – Nomnom Mar 31 '21 at 19:21
-
1I see the other answers are good debugging/troubleshooting info. I've tried to give the easiest and most direct way to find the deployment error. – 70Mike Nov 30 '21 at 19:35
-
Thanks @70Mike! – Nomnom Dec 02 '21 at 21:12
3 Answers
10
helm history <release_name>
Shows the kubernetes errors for the attempted deployment of that release.
C.J. Annunziato
- 3
- 3
70Mike
- 304
- 2
- 12
8
Helm chart deployment can fail because of the below 3 reasons.
- Chart is incorrect
- Deployment configurations are incorrect.
- Problem in Scheduling the containers and service due to some issue(image pull issue, resource quota)
The following steps can be performed for debugging.
Chart correctness check
Check the correctness of the chart using helm lint in local
helm lint PATH [flags] Ref: https://helm.sh/docs/helm/helm_lint/
Deployment configuration correctness check
Check the correctness of the proposed deployment configuration using the flag --dry-run --debug
helm install <ChartName> --dry-run --debug
Kubernetes cluster Deployment issue check
To check the issues such as whether container creation is failing due to image pull failure or the pods are not getting scheduled due to resource crunch perform the following diagnostic.
Please add the namespace flag in each of the commands if your deployment is getting deployed in a particular namespace ( append -n <yournamespace_name>). Look out for the items which are relevant to your helm chart.
kubectl get deployment
kubectl describe deployment <deployment_id_found_in_previos_step>
kubectl get rs
kubectl describe rs <resource_set_id_found_in_previos_step>
kubectl get pods
kubectl describe pods <pod_id_found_in_previos_step>
See the output of each of the above-mentioned commands.
Amit Meena
- 2,884
- 2
- 21
- 33
3
Append the below parameters to the helm install command to validate the syntax and the command
--dry-run --debug
if the command works, then go ahead and run skipping the above parameters.
P Ekambaram
- 15,499
- 7
- 34
- 59