Open yanchenzhang0304 opened 2 months ago
Increasing the OS disk size should be allowed, but decreasing it is not allowed.
As this would require each node to update, you can just create a new node pool with the correct size and then migrate the workloads in your own time to reduce any downtime.
I am not sure this feature would be needed. Can you give more information as to why you would need this feature?
As this would require each node to update, you can just create a new node pool with the correct size and then migrate the workloads in your own time to reduce any downtime.
I am not sure this feature would be needed. Can you give more information as to why you would need this feature?
Just like the feat https://github.com/Azure/AKS/issues/4485, we want to update the vmss configuration via in-place update rather than nodepool rotation. In-place changes have less impact on services, and the upgrade process is smoother (no need for users to implement graceful drain).
Provide an API to update nodepool OS disk size in place honoring PDBs and gradually updating the VMs.