We’re currently using the “Stopped OS disk type” setting to switch VM OS disks from Premium SSD to Standard HDD when the machines are stopped. This has generally worked as expected, but a recent outage exposed a gap in how this behaves when VMs are started outside of Nerdio automation.
In our environment:
- Some host pools also rely on a startup scripted action (Nerdio “Run script when VM is started”).
- During the outage, autoscale was not functioning, so the team started VMs directly from Azure.
- When this happens:
- The disk type does not get corrected (because the VM starts with whatever disk type it had previously).
- The startup scripted action does not run (since it wasn’t triggered through Nerdio).
- This resulted in:
- Hosts running on the wrong disk type
- Startup configuration steps being skipped
- Inconsistent host behavior across the pool
Proposed Enhancement
Add an option within autoscale configuration (under the OS disk type settings):
- A simple toggle/checkbox such as:
“If running disk type is incorrect, restart VM immediately”
- Optionally, add this same functionality to the Auto-heal settings
Expected Behavior:
- When a VM is detected as running with the wrong disk type:
- Nerdio automatically initiates a restart
- On restart:
- The correct OS disk type is applied
- The configured “run script on VM start” action is executed
Why this helps:
- Self-heals scenarios where VMs are started outside of Nerdio (Azure portal, manual recovery, etc.)
- Ensures consistency across host pools without requiring manual intervention
- Guarantees both:
- Correct disk performance/cost alignment
- Execution of required startup configuration scripts
- Reduces operational risk during outages or autoscale failures
Comments (2 comments)