Chassis only uses the Arm box on Arm if the mode is set to base

[tomjn]

If you run an Apple Silicon device and use parallels it will try to use the amd64/x86 box if the mode is set to normal.

There is a check to change it to the arm64 box, but this is only running when the mode is base:

	# The Parallels Provider uses a different naming scheme.
	config.vm.provider :parallels do |_v, override|
		unless CONF['_mode'] == "normal"
			# Vagrant currently runs under Rosetta on M1 devices. As a result,
			# this seems to be the most reliable way to detect whether or not we're
			# running under ARM64.
			if Etc.uname[:version].include? 'ARM64'
				override.vm.box = 'mpasternak/focal64-arm'
			end
		end
	end

instead, until an arm64 chassis box is built it should instead force the CONF['_mode'] value to base

[BronsonQuick]

This is by design at the moment and it's been documented.

Do you know if there's a way to detect in in the Vagrantfile? If we can detect it then I'm happy to add it. I don't have a Silicon chip so I can't test it but if you do a PR for us then I'll be happy to test the backwards compatibility and merge it in. Taa!

[tomjn]

I believe it's in the code block above, here is VVV's check to auto-switch our config to use the Parallels provider by default if Arm64 is detected as the chipset:

# if Arm default to parallels
if Etc.uname[:version].include? 'ARM64'
  defaults['provider'] = 'parallels'
end

[tomjn]

Perhaps the code I quoted can be simplified to:

	# The Parallels Provider uses a different naming scheme.
	config.vm.provider :parallels do |_v, override|
		# Vagrant currently runs under Rosetta on M1 devices. As a result,
		# this seems to be the most reliable way to detect whether or not we're
		# running under ARM64.
		if Etc.uname[:version].include? 'ARM64'
			override.vm.box = 'mpasternak/focal64-arm'
		end
	end

Then perhaps in def self.normalize_config(config) in puppet/chassis.rb it can do a check like this:

		if Etc.uname[:version].include? 'ARM64'
			config['_mode'] = 'base'
		end

What do you think?

[BronsonQuick]

@tomjn That looks great to me! Any chance you'd be able to pop up a PR for that and let me know if it works as expected for you then I'll test backwards compatibility on my Intel chip. Thanks!

[tomjn]

I've very very limited bandwidth available to do this, it may take a few weeks/months before I can get to this

Note that the code I shared is almost guaranteed to work though and is extremely similar to work done elsewhere

[BronsonQuick]

@tomjn Sorry for the epic delay on this. I'm doing some catching up now. I've added your fix. Hopefully it works for you. If you get a chance could you please do a little test for me? Thanks man!