DuplicateName exists: plugin cannot find current Powered On VMs using Multi-machine Loop

[ayen-tibco]

During my CI build, the .vagrant directory was accidentally delete while my VMs are still running on my ESXi, I need to rebuild the .vagrant and have vagrant up detect the current running VMs. But to no avail I can't get it running with my current multi-machine Vagrantfile. Any suggestions besides manually deleting my VMs and recreating them again?

vagrant version
Installed Version: 1.8.5
Latest Version: 1.8.5
vagrant-vsphere (1.10.0)
vSphere: 6.0U2

vagrant status
ERROR loader: Unknown config sources: ["21248980_machine_instance-1"]
ERROR loader: Unknown config sources: ["21248980_machine_instance-1"]
ERROR loader: Unknown config sources: ["21248980_machine_instance-1", "21248980_machine_instance-2"]
ERROR loader: Unknown config sources: ["21248980_machine_instance-1", "21248980_machine_instance-2"]
ERROR loader: Unknown config sources: ["21248980_machine_instance-2", "21248980_machine_instance-3"]
ERROR loader: Unknown config sources: ["21248980_machine_instance-1", "21248980_machine_instance-3"]
ERROR loader: Unknown config sources: ["21248980_machine_instance-1", "21248980_machine_instance-2"]
Current machine states:

instance-1 vagrant_vsphere.states.short_not_created (vsphere)
instance-2 vagrant_vsphere.states.short_not_created (vsphere)
instance-3 vagrant_vsphere.states.short_not_created (vsphere)

Vagrantfile:

Vagrant.require_version ">= 1.7.0"
cluster_machine_names=Array.new
unless Vagrant.has_plugin?("vsphere")
  raise 'vagrant-vsphere is not installed! Enter [vagrant plugin install vagrant-vsphere] to install.'
end

Vagrant.configure(2) do |config|
    # Disable the new default behavior introduced in Vagrant 1.7, to
    # ensure that all Vagrant machines will use the same SSH key pair.
    # See https://github.com/mitchellh/vagrant/issues/5005
    config.ssh.insert_key = false
    config.ssh.username = "vagrant"
    config.ssh.private_key_path = './example_box/vagrant_private.key'
    config.ssh.pty = true
    config.vm.box = "vsphere.box"
    config.vm.box_url = "example_box/vsphere.box"

    # Vsphere defaults
    config.vm.provider :vsphere do |vsphere|
        vsphere.host = 'somevcenterhost'
        vsphere.compute_resource_name = 'someesxihost'
        vsphere.resource_pool_name = ''
        vsphere.template_name = 'centos7'
        vsphere.user = 'vagrant@vsphere.local'
        vsphere.password = 'somepassword'
        # If you don't have SSL configured correctly, set this to 'true'
        vsphere.insecure = true
        vsphere.memory_mb = 4096
        vsphere.cpu_count = 2
    end
    N=3
    (1..N).each do |i|
        config.vm.define "instance-#{i}" do |instance|
            ins
https://guides.github.com/features/mastering-markdown/tance.vm.hostname = "instance-#{i}"
            cluster_machine_names << "instance-#{i}"
            instance.vm.network "private_network", type: "dhcp"
            # forwarded port doesn not work with vsphere plugin yet
            # instance.vm.network "forwarded_port", guest: 5556, host: host_admin_ports[N], auto_correct: true
            instance.vm.provider :vsphere do |vsphere|
                vsphere.name = "instance-#{i}"
            end

            # trick to get playbook run only once with parrallelism
            if i == N
                instance.vm.provision "ansible" do |ansible|
                    ansible.verbose = "v"
                    ansible.limit = "all"
                        # perform the entire deployment or custom tags
                        ansible.playbook = "playbooks/someplaybook.yml"
                    end
                end
           end
        end
    end
end

Debug Output:
vsphere.txt

[taliesins]

You might want the changes in #210 and then retry this.

[ayen-tibco]

Tested with #210 and doesn't resolve existings VMs created from prior vagrant up.

$ vagrant up A1
Bringing machine 'A1' up with 'vsphere' provider...
==> A1: Setting custom memory: 4096
==> A1: Setting custom cpu count: 2
Progress: 18%DuplicateName: The name 'A1' already exists.

$ vagrant status
Current machine states:

A1 not created (vsphere)
A2 not created (vsphere)

[taliesins]

Interesting. I don't think I affected anything around this.

I am thinking perhaps it's related to the state folder that is created for each vm by vagrant. I don't remember seeing anywhere a validation to see if the name already exists on vsphere.

Did you run vagrant up with this branch and then run it again and it failed? Or did you have a vm created with a previous branch and then used vagrant up.

I will look into it!