cpu_model gets replaced every time

[TGM]

Description

cpu_model gets replaced every time

Terraform and Provider version

Provider version: 1.2.1 OpenNebula version: 6.4.0

Affected resources and data sources

template_section / cpu_model

Terraform configuration

template_section {
    name     = "cpu_model"
    elements = {
      model = "host-passthrough"
    }

Expected behavior

  + template_section {
      + elements = {
          + "model" = "host-passthrough"
        }
      + name     = "cpu_model"
    }

cpu_model should not be replaced after provisioning

Actual behavior

cpu_model is replaced every time.

• cpumodel {
  • model = "host-passthrough" -> null }

Steps to Reproduce

Add the terraform configuration from above and try to provision a VM.

Debug output

No response

Panic output

No response

Important factoids

No response

References

No response

[treywelsh]

It seems cpumodel sections of the virtual machine resource is missing in the documentation, we need at least a PR to update the documentation.

See an usage example here: https://github.com/OpenNebula/terraform-provider-opennebula/blob/master/opennebula/resource_opennebula_virtual_machine_test.go#L605 Feel free to drop a comment if it works as expected, or not

[TGM]

If I use cpumodel as specified in the https://github.com/OpenNebula/terraform-provider-opennebula/blob/master/opennebula/resource_opennebula_virtual_machine_test.go#L605

The feature is added,

CPUMODEL = [
  MODEL = "host-passthrough" ]
DESCRIPTION = "Uses Proton default image that enables root login with a simple password; must not use outside of testing."
SCHED_REQUIREMENTS = "CLUSTER_ID=\"308\" "

but ignored, by OpenNebula.

[TGM]

Update: Works properly with

    model = "host-passthrough"
  }

but IMO should be renamed. Adding this PR here since I allready have it. https://github.com/OpenNebula/terraform-provider-opennebula/pull/454

TODO: add a docs PR regardless of the approach.

[TGM]

On the same page, setting the cpumodel to "" should have the same behavior as seting it to default. As far as I can tell default is "" but in the background opennebula assigns it to qemu64, even tough the GUI displays it as blank.

Currently setting it cpumodel to "" crashes the plugin.

  cpumodel {
    model = ""
  }

module.vmk["vmk-dev-intel-01"].netbox_primary_ip.primary_ip: Creation complete after 1s [id=1326]
╷
│ Error: Plugin did not respond
│ 
│   with module.vmk["vmk-dev-intel-01"].opennebula_virtual_machine.vm,
│   on ../single-vm/main.tf line 56, in resource "opennebula_virtual_machine" "vm":
│   56: resource "opennebula_virtual_machine" "vm" {
│ 
│ The plugin encountered an error, and failed to respond to the
│ plugin.(*GRPCProvider).ApplyResourceChange call. The plugin logs may
│ contain more details.
╵

Stack trace from the terraform-provider-opennebula_v1.2.1 plugin:

panic: interface conversion: interface {} is nil, not map[string]interface {}

goroutine 48 [running]:
github.com/OpenNebula/terraform-provider-opennebula/opennebula.generateVMTemplate(0xc68720?, 0xc000716708)
        github.com/OpenNebula/terraform-provider-opennebula/opennebula/shared_schemas.go:609 +0xae6
github.com/OpenNebula/terraform-provider-opennebula/opennebula.generateVm(0xc000326a68?, {0xc3f800?, 0xc000696ac0}, 0xc000350060)
        github.com/OpenNebula/terraform-provider-opennebula/opennebula/resource_opennebula_virtual_machine.go:2220 +0x34d
github.com/OpenNebula/terraform-provider-opennebula/opennebula.resourceOpennebulaVirtualMachineCreate({0xe81f48, 0xc0006921e0}, 0xc000128180, {0xc3f800?, 0xc000696ac0})
        github.com/OpenNebula/terraform-provider-opennebula/opennebula/resource_opennebula_virtual_machine.go:288 +0x305
github.com/hashicorp/terraform-plugin-sdk/v2/helper/schema.(*Resource).create(0xc0004a3b20, {0xe81f80, 0xc0005f87b0}, 0xd?, {0xc3f800, 0xc000696ac0})
        github.com/hashicorp/terraform-plugin-sdk/v2@v2.24.1/helper/schema/resource.go:707 +0x12e
github.com/hashicorp/terraform-plugin-sdk/v2/helper/schema.(*Resource).Apply(0xc0004a3b20, {0xe81f80, 0xc0005f87b0}, 0xc0005fb2b0, 0xc000128980, {0xc3f800, 0xc000696ac0})
        github.com/hashicorp/terraform-plugin-sdk/v2@v2.24.1/helper/schema/resource.go:837 +0xa7a
github.com/hashicorp/terraform-plugin-sdk/v2/helper/schema.(*GRPCProviderServer).ApplyResourceChange(0xc00000d728, {0xe81f80?, 0xc0005f8690?}, 0xc0005320f0)
        github.com/hashicorp/terraform-plugin-sdk/v2@v2.24.1/helper/schema/grpc_provider.go:1021 +0xe3c
github.com/hashicorp/terraform-plugin-go/tfprotov5/tf5server.(*server).ApplyResourceChange(0xc00009f4a0, {0xe81f80?, 0xc0005f8000?}, 0xc000626070)
        github.com/hashicorp/terraform-plugin-go@v0.14.1/tfprotov5/tf5server/server.go:818 +0x574
github.com/hashicorp/terraform-plugin-go/tfprotov5/internal/tfplugin5._Provider_ApplyResourceChange_Handler({0xd03360?, 0xc00009f4a0}, {0xe81f80, 0xc0005f8000}, 0xc000626000, 0x0)
        github.com/hashicorp/terraform-plugin-go@v0.14.1/tfprotov5/internal/tfplugin5/tfplugin5_grpc.pb.go:385 +0x170
google.golang.org/grpc.(*Server).processUnaryRPC(0xc0000003c0, {0xe85180, 0xc00050c820}, 0xc0005c8360, 0xc00048b6b0, 0x136f660, 0x0)
        google.golang.org/grpc@v1.50.1/server.go:1340 +0xd13
google.golang.org/grpc.(*Server).handleStream(0xc0000003c0, {0xe85180, 0xc00050c820}, 0xc0005c8360, 0x0)
        google.golang.org/grpc@v1.50.1/server.go:1713 +0xa1b
google.golang.org/grpc.(*Server).serveStreams.func1.2()
        google.golang.org/grpc@v1.50.1/server.go:965 +0x98
created by google.golang.org/grpc.(*Server).serveStreams.func1
        google.golang.org/grpc@v1.50.1/server.go:963 +0x28a

Error: The terraform-provider-opennebula_v1.2.1 plugin crashed!

This is always indicative of a bug within the plugin. It would be immensely
helpful if you could report the crash with the plugin's maintainers so that it
can be fixed. The output above should help diagnose the issue.

[TGM]

I'll close this one, and open a new one for what I've found.

[TGM]

Reopening, as this bug still remains

Setting model to "" does not revert back to default model.

  cpumodel {
    model = ""
  }

[github-actions[bot]]

This issue is stale because it has been open for 30 days with no activity and it has not the 'status: confirmed' label or it is not in a milestone. Remove the 'status: stale' label or comment, or this will be closed in 5 days.

[sk4zuzu]

It seems that CPU_MODEL->MODEL feature works correctly. The caveat is that we can't really modify cpu model in a running VM. Rebooting the VM is not enough, it must be powered down then resumed (by the user), only then vm.xml can be (and is) properly regenerated. So in the end, we allow such model change but it's not immediate. There is no auto-poweroff/on feature in OpenNebula, so we won't implement it in the provider either. :)