Need to simulate helm test failure while using KWOK.

[RomanBudnyk]

Hi all,

first of all - many thanks for developing and maintaining KWOK as a kube simulator, it is helping a lot!

I have a task - to simulate helm test failure and to see how my product will react to this. I was trying to achieve this by using Stages, but did not have any success. I tried this one https://github.com/kubernetes-sigs/kwok/blob/main/kustomize/stage/pod/chaos/pod-container-running-failed.yaml and different ones generated by AI or self-written, but each time pod is finishing with completed and not turning to an error state.

if you have any ideas how to help me with my tech task - it would be great.

I am suign KWOK v0.7.0, running inside docker.

thanks in advance!

[wzshiming]

May I know how you use it?

FYI To the best of my knowledge the current program relies on this, and I thought it would be a good reference for you.

KubeVirt
NVIDIA/knavigator

[RomanBudnyk]

yes, it is working fine.

the issue is that I want to move pods not to "completed" state, but to some kind of error state.

to achieve this, we've created a helm test and packed to the chart. for real cluster it would be enough to put exit code 1, but as KWOK does not create real containers, all my tries leads to succeed.

to verify some functionality of my product I need to somehow have failed pods, and this is main ask - how to achieve this.

thanks

[wzshiming]

https://github.com/NVIDIA/knavigator/blob/15be68c215a857eb1351afff40daa4d4c0f33ad2/resources/templates/k8s/failed-job.yml#L27-L35

Do you use it like this?

[RomanBudnyk]

@wzshiming

this is how helm is deploying this kind of pod:

Name: my-negative-test
Namespace: ns000-stability-nrm
Priority: 0
Service Account: default
Node: kwok-node-13/
Labels:
Annotations: helm.sh/hook: test
helm.sh/hook-delete-policy: before-hook-creation
helm.sh/hook-weight: 0
kwok.x-k8s.io/chaos: pod-container-failed
Status: Succeeded
IP: 10.0.0.16
IPs:
IP: 10.0.0.16
Containers:
negative-test:
Container ID:
Image: busybox
Image ID:
Port:
Host Port:
Command:
sh
-c
echo "This test will always fail"; exit 1
State: Terminated
Reason: Completed
Exit Code: 0
Started: Tue, 30 Sep 2025 18:25:23 +0000
Finished: Tue, 30 Sep 2025 18:25:23 +0000
Ready: True
Restart Count: 0
Environment:
Mounts:
/var/run/secrets/kubernetes.io/serviceaccount from kube-api-access-65wzb (ro)
Conditions:
Type Status
Initialized True
Ready True
ContainersReady True
PodScheduled True
Volumes:
kube-api-access-65wzb:
Type: Projected (a volume that contains injected data from multiple sources)
TokenExpirationSeconds: 3607
ConfigMapName: kube-root-ca.crt
Optional: false
DownwardAPI: true
QoS Class: BestEffort
Node-Selectors:
Tolerations: node.kubernetes.io/not-ready:NoExecute op=Exists for 300s
node.kubernetes.io/unreachable:NoExecute op=Exists for 300s
Events:

I do not see that it is creating "jobs" there

[wzshiming]

So who implemented the kwok.x-k8s.io/chaos,

In the case of knavigator it is associated with the label pod-init-container-running-failed.stage.kwok.x-k8s.io, and where is defined in the Stage CR

[RomanBudnyk]

I see, where are you going. let me try this, will return to you. thanks!

[RomanBudnyk]

I achieved what I needed. thanks for your help!