You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
What happened:
I use pod-delete experiment and my parent resource definition is custom (StrimziPodSet in my case)
So when experiment starts - it deletes pod, but then failes to verify recovering
Here are some parameters (i tried to play with different values to get it work, but failed):
I tried to use "statefulset:kafka-external:odf-cluster-kafka", "statefulset:kafka-external:odf-cluster-kafka-0" as TARGETS (and other variants), but it fails What you expected to happen:
Verification of pod recovering should work
Here is a full log from helper pod:
time="2024-04-30T10:18:06Z" level=info msg="Experiment Name: pod-delete"
time="2024-04-30T10:18:06Z" level=info msg="[PreReq]: Getting the ENV for the pod-delete experiment"
time="2024-04-30T10:18:08Z" level=info msg="[PreReq]: Updating the chaos result of pod-delete experiment (SOT)"
time="2024-04-30T10:18:10Z" level=info msg="The application information is as follows" Targets="[{namespace: kafka-external, kind: strimzipodset, names: [odf-cluster-kafka]}]" Chaos Duration=50
time="2024-04-30T10:18:12Z" level=info msg="[Info]: The chaos tunables are:" PodsAffectedPerc=0 Sequence=parallel
time="2024-04-30T10:18:14Z" level=info msg="[Chaos]:Number of pods targeted: 1"
time="2024-04-30T10:18:14Z" level=info msg="Target pods list for chaos, [odf-cluster-kafka-0]"
time="2024-04-30T10:18:14Z" level=info msg="[Info]: Killing the following pods" PodName=odf-cluster-kafka-0
time="2024-04-30T10:18:14Z" level=info msg="[Wait]: Wait for the chaos interval 10s"
time="2024-04-30T10:18:24Z" level=info msg="[Status]: Verification for the recreation of application pod"
time="2024-04-30T10:18:24Z" level=error msg="Chaos injection failed, err: could not run chaos in parallel mode\n --- at /litmus-go/chaoslib/litmus/pod-delete/lib/pod-delete.go:50 (PreparePodDelete) ---\nCaused by: could not check pod statuses by workload names\n --- at /litmus-go/chaoslib/litmus/pod-delete/lib/pod-delete.go:244 (injectChaosInParallelMode) ---\nCaused by: could not get pods by workload names\n --- at /litmus-go/pkg/status/application.go:330 (CheckUnTerminatedPodStatusesByWorkloadName) ---\nCaused by: {"errorCode":"TARGET_SELECTION_ERROR","reason":"no pod found for specified target","target":"{namespace: kafka-external, kind: , name: odf-cluster-kafka}"}"
As you can see - "kind" is empty
The text was updated successfully, but these errors were encountered:
This is a BUG report
What happened:
I use pod-delete experiment and my parent resource definition is custom (StrimziPodSet in my case)
So when experiment starts - it deletes pod, but then failes to verify recovering
Here are some parameters (i tried to play with different values to get it work, but failed):
metadata: name: kafka-pod-delete namespace: kafka-external spec: engineState: active annotationCheck: "false" appinfo: appns: "kafka-external" applabel: "statefulset.kubernetes.io/pod-name=odf-cluster-kafka-0" appkind: "statefulset" chaosServiceAccount: general-sa experiments: - name: pod-delete spec: components: env: - name: TOTAL_CHAOS_DURATION value: "50" - name: TARGETS value: "strimzipodset:kafka-external:odf-cluster-kafka" - name: TARGET_PODS value: "odf-cluster-kafka-0" - name: POD_NAME value: "odf-cluster-kafka-0" - name: CHAOS_NAMESPACE value: "kafka-external"
I tried to use "statefulset:kafka-external:odf-cluster-kafka", "statefulset:kafka-external:odf-cluster-kafka-0" as TARGETS (and other variants), but it fails
What you expected to happen:
Verification of pod recovering should work
What I see in code:
https://github.com/litmuschaos/litmus-go/blob/master/pkg/workloads/workloads.go#L73
In this method it tried to get pod owner type and name. But if it's a custom "Kind" - it will return an empty "parentType"
And after that here:
https://github.com/litmuschaos/litmus-go/blob/master/pkg/workloads/workloads.go#L58
it will fail as "ownerType" is empty
Here is a full log from helper pod:
time="2024-04-30T10:18:06Z" level=info msg="Experiment Name: pod-delete"
time="2024-04-30T10:18:06Z" level=info msg="[PreReq]: Getting the ENV for the pod-delete experiment"
time="2024-04-30T10:18:08Z" level=info msg="[PreReq]: Updating the chaos result of pod-delete experiment (SOT)"
time="2024-04-30T10:18:10Z" level=info msg="The application information is as follows" Targets="[{namespace: kafka-external, kind: strimzipodset, names: [odf-cluster-kafka]}]" Chaos Duration=50
time="2024-04-30T10:18:12Z" level=info msg="[Info]: The chaos tunables are:" PodsAffectedPerc=0 Sequence=parallel
time="2024-04-30T10:18:14Z" level=info msg="[Chaos]:Number of pods targeted: 1"
time="2024-04-30T10:18:14Z" level=info msg="Target pods list for chaos, [odf-cluster-kafka-0]"
time="2024-04-30T10:18:14Z" level=info msg="[Info]: Killing the following pods" PodName=odf-cluster-kafka-0
time="2024-04-30T10:18:14Z" level=info msg="[Wait]: Wait for the chaos interval 10s"
time="2024-04-30T10:18:24Z" level=info msg="[Status]: Verification for the recreation of application pod"
time="2024-04-30T10:18:24Z" level=error msg="Chaos injection failed, err: could not run chaos in parallel mode\n --- at /litmus-go/chaoslib/litmus/pod-delete/lib/pod-delete.go:50 (PreparePodDelete) ---\nCaused by: could not check pod statuses by workload names\n --- at /litmus-go/chaoslib/litmus/pod-delete/lib/pod-delete.go:244 (injectChaosInParallelMode) ---\nCaused by: could not get pods by workload names\n --- at /litmus-go/pkg/status/application.go:330 (CheckUnTerminatedPodStatusesByWorkloadName) ---\nCaused by: {"errorCode":"TARGET_SELECTION_ERROR","reason":"no pod found for specified target","target":"{namespace: kafka-external, kind: , name: odf-cluster-kafka}"}"
As you can see - "kind" is empty
The text was updated successfully, but these errors were encountered: