# Ansible Playbook Running Against Wrong Vagrant VM

**URL:** <https://forum.ansible.com/t/ansible-playbook-running-against-wrong-vagrant-vm/21417>\
**Category:** Ansible Project\
**Created:** [May 11, 2016, 9:57pm UTC](https://forum.ansible.com/t/ansible-playbook-running-against-wrong-vagrant-vm/21417 "2016-05-11T21:57:36Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Robert\_F](https://avatars.discourse-cdn.com/v4/letter/r/9d8465/32.png) [@Robert\_F](https://forum.ansible.com/u/Robert_F)\
**Post date:** [May 11, 2016, 9:57pm UTC](https://forum.ansible.com/t/ansible-playbook-running-against-wrong-vagrant-vm/21417/1 "2016-05-11T21:57:36Z")

</div>

I have an Ansible playbook which, when I run it, executes its tasks against the wrong Vagrant VM. Here’s how I built my VMs:

# ~/playbooks/vagrant/Vagrantfile

Vagrant.configure(2) do |config|  
config.ssh.insert\_key = false  
config.vm.box = “vagrant\_base” # Fresh debian/jessie64 image  
config.vm.define “web” do |web|  
web.vm.network “private\_network”, ip: “192.168.2.100”  
end  
config.vm.define “db” do |db|  
db.vm.network “private\_network”, ip: “192.168.2.101”  
end  
config.vm.define “fs” do |fs|  
fs.vm.network “private\_network”, ip: “192.168.2.102”  
end  
end

I run the playbook with this command:

ansible-playbook -i inventories/vagrant site.yml

The playbook just calls playbooks for each of my servers:

# ~/playbooks/example/site.yml

---

<div class="post-metadata">

**Author:** ![Robert\_F](https://avatars.discourse-cdn.com/v4/letter/r/9d8465/32.png) [@Robert\_F](https://forum.ansible.com/u/Robert_F)\
**Post date:** [May 11, 2016, 10:15pm UTC](https://forum.ansible.com/t/ansible-playbook-running-against-wrong-vagrant-vm/21417/2 "2016-05-11T22:15:18Z")

</div>

I just noticed that if I only bring the file server VM up and run the ansible setup command, I get an SSH error saying, “data could not be sent to the remote host”. But if I bring the database VM up, the command against the file server runs but the ansible\_hostname is “db00”, not “fs00” as I would expect. Is this ansible\_hostname fact what Ansible looks at when it tries to identify what server to run the playbook against? That’s what it seems like. I think this may be related to the fact that I created the Vagrant VMs in one directory (playbooks/vagrant) while my actual Ansible playbooks reside in another (playbooks/example). I did this because I initially wrote my playbooks to run against real Debian servers. They run just fine against those servers. I’m only having problems running them against my Vagrant VMs which I built later.

One thing I’m confused about is ansible\_hostname versus inventory\_hostname. When should each of them be used?

By the way, I’m just using [example.com](http://example.com) as an example. It’s not my real domain name.

---

<div class="post-metadata">

**Author:** ![Robert\_F](https://avatars.discourse-cdn.com/v4/letter/r/9d8465/32.png) [@Robert\_F](https://forum.ansible.com/u/Robert_F)\
**Post date:** [May 12, 2016, 10:15pm UTC](https://forum.ansible.com/t/ansible-playbook-running-against-wrong-vagrant-vm/21417/3 "2016-05-12T22:15:21Z")

</div>

The problem was that I needed to use the inventory file that Vagrant creates in .vagrant/provisioners/ansible/inventory/vagrant\_ansible\_inventory instead of the inventory/vagrant file I had created. I contains the necessary Ansible directives.