FailedAttachVolume with K8s backend in GKE Autopilot

[xarlybovi]

Component

agent

Describe the bug

Due to the nature of the GKE Autopilot cluster and my budget limit I have to configure the storage as RWO. The problem is that several pods can potentially be created at different nodes. This causes the volume not to be released until the pod of the corresponding step is finished.
In case is a service pod that does not terminate until the end of the workflow it can have the volume forever and end up giving timeout.
Would it be possible not to mount volume in the services?

Steps to reproduce

1. Install Woodpecker CI server and kubernetes backend agents with the helm chart.
2. Run a repo pipeline with parallel steps and a service.

Expected behavior

I would like the workflow not to get stuck when a service attaches the volume. We could avoid this by making it optional to mount the volume on the service.

System Info

Woodpecker CI 3.0.1
GKE Autopilot 1.30.8-gke

Additional context

No response

Validations

• Read the docs.
• Check that there isn't already an issue that reports the same bug to avoid creating a duplicate.
• Checked that the bug isn't fixed in the next version already [https://woodpecker-ci.org/faq#which-version-of-woodpecker-should-i-use]