sandbox provisioning timed out after 300s.

[mcrisafu]

Hi everyone,

First of all, thank you for all the effort put into this project. It's really impressive.

I'm running OpenShell on an AKS cluster and followed the installation guide, including the Ingress setup. The gateway is accessible and I can create sandbox resources, but sandbox provisioning (in the CLI) never completes.

When running:

openshell sandbox create -- echo "test"

the command times out after 300 seconds with:

✗ sandbox provisioning timed out after 300s. Last reported status: DependenciesNotReady: Pod exists with phase: Pending
✓ Sandbox allocated (7s)
✓ Sandbox allocated (7s)
✓ Image pulled (9s)

Error:
× sandbox provisioning timed out after 300s. Last reported status: DependenciesNotReady: Pod exists with phase: Pending

Listing the sandbox shows:

❯ openshell sandbox list
NAME                CREATED              PHASE
masterful-angora    2026-06-11 19:50:12  Unspecified

I checked the logs and don't see any obvious errors. The only warnings I noticed are of this type:

WARN openshell_sandbox::opa: Cannot access container filesystem for symlink resolution: path=/usr/bin/wget

Has anyone encountered this behavior before?

Additional environment details:

• OpenShell version: 0.0.59
• OpenShell CLI version: 0.0.44
• Kubernetes version: 1.34.7

The pod is created successfully and works (when getting a shell via kubectl).

Thanks in advance for any suggestions.

[Advait251206]

The fact that the pod is created successfully and is usable via kubectl exec suggests the issue is likely not with the sandbox container itself, but rather with whatever readiness condition OpenShell is waiting for before marking the sandbox as provisioned.

A few things stand out:

• Your OpenShell server (0.0.59) and CLI (0.0.44) are on different versions. While this may not be the root cause, I'd recommend upgrading the CLI to match the server version (or vice versa), as provisioning protocols can change between releases.
• The status repeatedly reports:

  DependenciesNotReady: Pod exists with phase: Pending
  

  even though you mention the pod is actually running and you can interact with it. That makes me wonder whether OpenShell's controller is failing to observe the pod's updated status rather than the pod actually remaining in Pending.
• The warning about /usr/bin/wget doesn't immediately look related. It appears to be an OPA/symlink resolution warning and, by itself, wouldn't normally explain a provisioning timeout.

I'd suggest checking a few things:

• Does kubectl get pod <sandbox-pod> -o wide report the pod as Running and Ready, or does Kubernetes still consider it Pending?
• Does kubectl describe pod <sandbox-pod> show any failing readiness probes, volume mounts, or scheduling events?
• Are the controller/operator logs showing repeated attempts to reconcile the sandbox, or are they stuck waiting for a condition that never becomes true?

Since the pod itself appears functional, this feels more like a mismatch between the actual pod state and what the OpenShell controller is expecting.

It would also be useful if the maintainers could clarify:

1. Are OpenShell 0.0.59 and CLI 0.0.44 expected to be compatible?
2. What exact Kubernetes condition or readiness check does the provisioning logic wait for before transitioning the sandbox out of DependenciesNotReady?
3. Is this a known issue on AKS or Kubernetes 1.34.x?

Those answers would help determine whether this is a version compatibility issue, an AKS-specific behavior, or a bug in the provisioning controller.

---

If this answer helped or pointed you in the right direction, I'd appreciate it if you could mark it as the accepted answer so it's easier for others with the same issue to find.

Also, if you found my contribution useful, I'd appreciate it if you could check out my GitHub profile, follow me, and star any repositories you find interesting.

GitHub: https://github.com/Advait251206