Run Command stuck on "execution is in progress" for 30h; VMAccessForLinux extension stuck in Creating

Question

Run Command stuck on "execution is in progress" for 30h; VMAccessForLinux extension stuck in Creating

Luan Vo 0

Linux VM, Malaysia West. Every az vm run-command call fails with:

(Conflict) Run command extension execution is in progress. Please wait for completion before invoking a run command.

30+ hours now. Nothing is actually running.

State:

VM provisioningState cycles between Updating and Failed/VMStartTimedOut. No PowerState entry at all in the instance view.
The VM has one extension, enablevmaccess (Microsoft.OSTCExtensions, VMAccessForLinux 1.5), stuck in provisioningState Creating with no instanceView.
Six managed runCommands objects stuck in Deleting.
All six show executionState Pending with startTime null, so none ever started. They are queued, not hung.

Guest is healthy and was never stopped:

Uptime 21 days continuous, no reboot
No run-command process running
No RunCommand or VMAccessForLinux directory under /var/lib/waagent
The run-command handler logged a clean uninstall 45 minutes BEFORE the VMStartTimedOut error
All services on the box are up and serving traffic

Trigger was adding a third ipConfig to the VM's NIC. That succeeded.

Tried, all accepted then stalled with no terminal state: deleting the six runCommands objects, az vm restart (exit 0, guest never rebooted), az vm start, az vm extension delete on enablevmaccess, tagging to force a reconcile. No cancel operation exists on either resource type.

The Portal troubleshooter ran and reported no actionable root cause.

Not attempting stop/deallocate: shared box, and a database container on it is restart=no.

Is the stuck enablevmaccess extension what is blocking Run Command?

Is there any way to clear it that I have missed?

If this needs backend cleanup, can it be escalated - my subscription is Developer plan so I cannot open a technical ticket directly.

Can supply subscription ID, VM resource ID and correlation IDs privately.

Hemant Bharadwaj 90 Reputation points Microsoft Employee

2026-11-03T11:34:51.69+00:00
Likely cause/summary: yes — your evidence strongly suggests the stuck extension state in control plane is blocking new Run Command executions (the platform still sees an in-progress extension operation even though guest-side handlers are gone).

Your signals are consistent with this:

Run command extension execution is in progress for 30+ hours,

VMAccessForLinux extension stuck in Creating with no instance view,

runCommand child resources stuck in Deleting and never started,

guest OS healthy and no active run-command process.

Validate VM agent and extension handler health from guest.

Ensure Azure Linux VM agent is running and healthy.

Collect waagent logs and extension-handler logs for the stuck window.

Run Command relies on the VM agent and extension pipeline.

Reference: https://learn.microsoft.com/en-us/azure/virtual-machines/linux/run-command

Force extension model convergence in ARM (safe-first order).

Re-GET extension/resource state, then attempt explicit delete/update of the stuck VMAccessForLinux extension.

Re-GET VM model + instanceView after each operation and wait for terminal provisioning state.

Avoid parallel extension operations while the model is inconsistent.

Use VM Assist for Linux to capture diagnostics package.

This gives support-ready evidence for agent/extension drift quickly.

Reference: https://learn.microsoft.com/en-us/azure/virtual-machines/extensions/features-linux

If still stuck, this typically needs backend cleanup.

Ask Microsoft support/moderator to escalate with:

subscription ID, VM resource ID

correlation IDs for failed run-command and extension operations

timestamps (UTC)

output showing runCommands stuck in Deleting + extension in Creating

Risk/downtime note.

A stop/deallocate or host-level maintenance action may clear control-plane drift but can impact workload continuity.

Since this is a shared box with strict uptime constraints, escalation-first is reasonable.

Confirmed facts from your post:

Guest is up and serving traffic.

No active run-command process exists guest-side.

Control plane remains blocked on extension/runCommand lifecycle.

Possible cause (not yet confirmed): stale extension orchestration state in Azure control plane after NIC update/VM timeout event.

Validation checks after remediation:

VM extension provisioning reaches terminal state.

az vm run-command invoke returns successfully.

No residual runCommand resources stuck in Deleting/Pending.

“If you find this answer helpful, please upvote it. If this resolves your issue, please mark it as the accepted answer so that it can help others in the community.”

Run Command stuck on "execution is in progress" for 30h; VMAccessForLinux extension stuck in Creating

Your answer