[BUG] disk_size_gb hardcoded in /usr/local/epicli/data/azure/defaults/infrastructure/virtual-machine.yml

[wkalbar]

NOTE: See comment from @przemyslavic below, should be a simple backport to add this 1.0.3

Describe the bug
Variable disk_size_gb for VMs OS disks used in config.yml is overwritten by variable in /usr/local/epicli/data/azure/defaults/infrastructure/virtual-machine.yml

How to reproduce
Steps to reproduce the behavior:

1. execute epicli init ... (with params)
2. edit config file - change variable disk_size_gb to any value other than default 32GB
3. execute epicli apply ...

Expected behavior
VMs should have OS disk size equal to value of disk_size_gb variable. Instead of this they have still 32Gb OS disks.

Config files
n/a

Environment

• Cloud provider: Azure
• OS: Ubuntu 18.046 LTS

epicli version: 1.0.2

Additional context
There is workaround to change this variable in /usr/local/epicli/data/azure/defaults/infrastructure/virtual-machine.yml inside container. But it's still workaround.

---

DoD checklist

• Changelog
  • updated
  • not needed
• COMPONENTS.md
  • updated
  • not needed
• Schema
  • updated
  • not needed
• Backport tasks
  • created
  • not needed
• Documentation
  • added
  • updated
  • not needed
• Feature has automated tests
• Automated tests passed (QA pipelines)
  • apply
  • upgrade
  • backup/restore
• Idempotency tested
• All conversations in PR resolved

[seriva]

Hey @wkalbar,

Did you define it properly inside your input config:

...
---
kind: infrastructure/virtual-machine
title: "Virtual Machine Infra"
provider: azure
name: default
version: 1.0.2
specification:
  storage_os_disk:
    disk_size_gb: 64 #for example
---
...

Pls add your config yaml so we can have a look what might be going on.

[wkalbar]

grep -B 5 disk_size_gb kilo-22061.yml

  storage_os_disk:
    delete_on_termination: false
    managed: false
    caching: ReadWrite
    create_option: FromImage
    disk_size_gb: 64
--
title: Virtual Machine Infra
provider: azure
name: repository-machine
specification:
  storage_os_disk:
    disk_size_gb: 64

[przemyslavic]

What works for me is:

...    
  components:
    repository:
      count: 1
    kubernetes_master:
      count: 1
    kubernetes_node:
      count: 1
    logging:
      count: 0
    monitoring:
      count: 0
    kafka:
      count: 0
    postgresql:
      count: 0
    load_balancer:
      count: 0
    rabbitmq:
      count: 0
version: 1.0.2
---
kind: infrastructure/virtual-machine
name: kubernetes-master-machine
provider: azure
based_on: kubernetes-master-machine
specification:
  storage_os_disk:
    disk_size_gb: 128
---
kind: infrastructure/virtual-machine
name: kubernetes-node-machine
provider: azure
based_on: kubernetes-node-machine
specification:
  storage_os_disk:
    disk_size_gb: 256

If you have a full configuration (generated from init --full) you have to attach:

  storage_os_disk:
    disk_size_gb: xxx

to the spec for each machine you want to change the disk size for.

Nevertheless, I confirm that changing the disk in only one place for the default machine is not working properly. It works fine starting form v1.1.0.

Comparing v1.0 to v1.1:
https://github.com/epiphany-platform/epiphany/blob/v1.0/core/src/epicli/cli/engine/schema/DefaultMerger.py
https://github.com/epiphany-platform/epiphany/blob/v1.1/core/src/epicli/cli/engine/schema/DefaultMerger.py
This has been added:

        default_config = select_first(self.docs, lambda x: x.name == 'default' and x.kind == doc.kind)

        if default_config is not None:
            merge_objdict(default_doc, default_config)

I think this could be the reason.

[seriva]

@przemyslavic Yes you are right, we didn't backport this apparently. Will mark it as backport for 1.0.3 and will fix it.