Support for automatic_instance_repair on the Orchestrated Virtual Machine Scale Set azure rm resource

[githubofkrishnadhas]

Is there an existing issue for this?

• I have searched the existing issues

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment and review the contribution guide to help.

Description

Hi Team,

I am using azurerm_orchestrated_virtual_machine_scale_set to spin up Axure VMSS using terraform.

The reason why i am using azurerm_orchestrated_virtual_machine_scale_set is that i need to test and try out the spot mix property introduced by Microsoft which enables us to have on demand as well as spot instances in same virtual machine scaleset.

I was using azurerm_linux_virtual_machine_scale_set resource and saw that this will not support spot mix, and only support uniform orchestration mode as either completely ondemand or spot .

I was unaware of azurerm_orchestrated_virtual_machine_scale_set module and raised the below issue: #24184

and from this i came to know about this resouce block.

the issue is that i need to use automatic_instance_repair as we will be using docker to set up the application and in case the containers went down this is a feature that can help us to an extent.

I saw this warning in terraform azurerm documentation:

To enable the automatic_instance_repair, the Orchestrated Virtual Machine Scale Set must have a valid health_probe_id

and for this to work we need to have a upgrade_mode and health_probe_id. but i cant see where to put it as its not clear in the documentation and code.

I can see comments saying this is not supported in ovmss

• https://github.com/hashicorp/terraform-provider-azurerm/blob/main/internal/services/compute/orchestrated_virtual_machine_scale_set_resource.go#L339-L341

This could be very helpful if you can integrate it also with OVMSS.

New or Affected Resource(s)/Data Source(s)

azurerm_orchestrated_virtual_machine_scale_set

Potential Terraform Configuration

This is how its in azurerm_linux_virtual_machine_scale_set 

resource "azurerm_linux_virtual_machine_scale_set" "vmss" {
  name                = azurerm_resource_group.vmss.name
  location            = azurerm_resource_group.vmss.location
  resource_group_name = azurerm_resource_group.vmss.name
  #tags                = azurerm_resource_group.vmss.tags
  depends_on = [
    azurerm_lb_rule.vmss
  ]
  upgrade_mode = "Manual"
  health_probe_id = "${ var.auto_repair ? azurerm_lb_probe.vmss.id : null}"
  zones = var.zones
  sku = "${var.sku}"
  instances = "${var.autoscale_default}"

// rest of terraform code
}

References

https://registry.terraform.io/providers/hashicorp/azurerm/latest/docs/resources/orchestrated_virtual_machine_scale_set

https://github.com/hashicorp/terraform-provider-azurerm/blob/main/internal/services/compute/orchestrated_virtual_machine_scale_set_resource.go#L339-L341

#24184
https://learn.microsoft.com/en-us/azure/virtual-machine-scale-sets/spot-priority-mix?tabs=portal

[harshavmb]

Hi @githubofkrishnadhas ,

Upgrade Mode policy for Flex orchestration mode is currently in preview mode. To activate it you could use below az-cli command ..

az feature registration create --subscription <sub-name/id> -n VmssFlexRollingUpgrade --namespace Microsoft.Compute

Having spoken to MS product team working on this feature, I am told that this is expected to be GA by the end Q1 2024. I don't think azurerm provider should support this for a private preview functionality as preview features are quite unstable.

MS product team told me that they will update when this feature is GA, then we could consider.

Disclaimer :: I am also one of the heavy patronizers of vmss flex mode, so I would definitely like to have this feature but only when API is stable.

[githubofkrishnadhas]

Hi @harshavmb Thanks for the update. Will wait for this to be generally available. :)

[darren-johnson]

Is there any update on this issue? I have just had to log 28270 in respect to rolling_upgrade_policy being missing too?