Not able to do packet capture, fails with "Error: the server does not allow this method on the requested resource"

[SohamChakraborty]

Describe the bug
I am trying to do a packet capture. It fails with the following error:

ts=2025-05-05T22:54:11.485+0530 level=error caller=capture/create.go:133 msg="Failed to create job" error="the server does not allow this method on the requested resource"
Error: the server does not allow this method on the requested resource

To Reproduce

$ kubectl retina capture create --pod-selectors="app=foobar"
ts=2025-05-05T22:54:09.013+0530 level=info caller=capture/create.go:259 msg="The capture duration is set to 1m0s"
ts=2025-05-05T22:54:09.013+0530 level=info caller=capture/create.go:266 msg="Overriding default node selectors value and setting it to nil. Using namespace and pod selectors. To use node selector, please remove namespace and pod selectors."
ts=2025-05-05T22:54:09.013+0530 level=info caller=capture/create.go:313 msg="The capture file max size is set to 100MB"
ts=2025-05-05T22:54:09.013+0530 level=info caller=utils/capture_image.go:57 msg="Using capture workload image ghcr.io/microsoft/retina/retina-agent:v0.0.32 with version determined by CLI version"
ts=2025-05-05T22:54:09.015+0530 level=info caller=capture/crd_to_job.go:202 msg="HostPath is not empty" HostPath=/mnt/retina/captures
ts=2025-05-05T22:54:11.235+0530 level=info caller=capture/crd_to_job.go:902 msg="The Parsed tcpdump filter is \"\""
#########################
Expected Capture Files
#########################
retina-capture-i-1234-20250505172411UTC.tar.gz
retina-capture-i-4321-20250505172411UTC.tar.gz

Note: The file(s) may not be created if the capture job(s) fail prematurely.
#########################
ts=2025-05-05T22:54:11.485+0530 level=error caller=capture/create.go:133 msg="Failed to create job" error="the server does not allow this method on the requested resource"
Error: the server does not allow this method on the requested resource
Usage:
  kubectl-retina capture create [flags]

Expected behavior
Packet capture should happen.

Platform (please complete the following information):

• OS: [e.g. AzureLinux] Ubuntu 22.04
• Kubernetes Version: [e.g. 1.22] 1.29
• Host: [e.g. AKS, KIND, self-host, etc] self hosted with kOps
• Retina Version: v0.0.32

FWIW, I ran the same command with -debug and I see same result.

[nddq]

seems like an RBAC issue, can you check the clusterrole to see if Retina has the right to create the capture resource?

[dezsibalint]

Hi, i have the same issue, I checked the clusterrole it looks fine to me. Any other idea what could be the problem?

[nddq]

@kamilprz can you help taking a look at this issue?

[kamilprz]

@SohamChakraborty / @dezsibalint have you tried to run pod-selectors coupled with namespace-selectors?

e.g. https://retina.sh/docs/Captures/cli#target-selection

[SohamChakraborty]

Hi @kamilprz ,
Sorry for being MIA.

I tested just now:

$ kubectl retina capture create   --name example-pod-namespace-selectors   --pod-selectors="app=metrics-receiver"   --namespace-selectors="kubernetes.io/metadata.name=default"
ts=2025-05-18T11:47:51.419+0530 level=info caller=capture/create.go:259 msg="The capture duration is set to 1m0s"
ts=2025-05-18T11:47:51.419+0530 level=info caller=capture/create.go:266 msg="Overriding default node selectors value and setting it to nil. Using namespace and pod selectors. To use node selector, please remove namespace and pod selectors."
ts=2025-05-18T11:47:51.419+0530 level=info caller=capture/create.go:313 msg="The capture file max size is set to 100MB"
ts=2025-05-18T11:47:51.419+0530 level=info caller=utils/capture_image.go:57 msg="Using capture workload image ghcr.io/microsoft/retina/retina-agent:v0.0.32 with version determined by CLI version"
ts=2025-05-18T11:47:51.421+0530 level=info caller=capture/crd_to_job.go:202 msg="HostPath is not empty" HostPath=/mnt/retina/captures
ts=2025-05-18T11:47:53.279+0530 level=info caller=capture/crd_to_job.go:902 msg="The Parsed tcpdump filter is \"\""
#########################
Expected Capture Files
#########################
example-pod-namespace-selectors-i-1234-20250518061753UTC.tar.gz
example-pod-namespace-selectors-i-4321-20250518061753UTC.tar.gz

Note: The file(s) may not be created if the capture job(s) fail prematurely.
#########################
ts=2025-05-18T11:47:53.477+0530 level=error caller=capture/create.go:133 msg="Failed to create job" error="the server does not allow this method on the requested resource"
Error: the server does not allow this method on the requested resource
Usage:
  kubectl-retina capture create [flags]
<SNIPPED>

I tried with --debug as well but that also shows me the same lines.

I am open to join any dev/users call or office hours if you are interested.

[kamilprz]

Thanks for that. I will try to take a look within the next week - have a couple of other issues which need my attention first :)