CH04 ansible-haproxy : configure haproxy service fails

[ful09003]

Replicated a few times now: TASK [ansible-haproxy : configure haproxy service] ***** fatal: [142.93.83.70]: FAILED! => {"changed": false, "msg": "AnsibleUndefinedVariable: 'ansible_eth0_2' is undefined"} fatal: [142.93.83.63]: FAILED! => {"changed": false, "msg": "AnsibleUndefinedVariable: 'ansible_eth0_2' is undefined"}

PLAY RECAP ***** 142.93.83.63 : ok=35 changed=33 unreachable=0 failed=1 142.93.83.64 : ok=1 changed=0 unreachable=0 failed=0 142.93.83.69 : ok=1 changed=0 unreachable=0 failed=0 142.93.83.70 : ok=35 changed=33 unreachable=0 failed=1 142.93.90.3 : ok=1 changed=0 unreachable=0 failed=0 159.89.220.217 : ok=1 changed=0 unreachable=0 failed=0

Looks like ansible_eth0_2 really isn't defined: ansible -i /usr/local/bin/terraform-inventory all -m setup --tree /tmp/facts > /tmp/ansible-debug grep -C 3 ansible_eth0_2 /tmp/ansible-debug (empty) grep -C 3 ansible_eth0_1 /tmp/ansible-debug (lots of hits for Droplet private IPs)

Steps to replicate:

1. Proceed through navguide to this part of ch04
2. Attempt to run the suggested ansible-playbook -i /usr/local/bin/terraform-inventory site.yml
3. Observe failure

[ful09003]

Adding another task in roles/ansible-haproxy/tasks/main.yml to register the Droplet's Anchor IP from the DigitalOcean metadata service, and using that in the jinja template will resolve this, and allows us to continue through the chapter.

[ful09003]

Seeing this in ch05 as well (task for ansible-galera-lb : configure haproxy service for galera)

[ful09003]

Re-tested with Ansible 2.5.8, and can confirm this behavior still exists:

ansible-playbook --version
ansible-playbook 2.5.8 (stable-2.5 36fd4a2255) last updated 2018/08/29 16:48:46 (GMT +000)
  config file = /root/navigators-guide/example-code/02-scale/ch04/haproxy_loadbalancer/ansible.cfg
  configured module search path = [u'/root/.ansible/plugins/modules', u'/usr/share/ansible/plugins/modules']
  ansible python module location = /root/ansible/lib/ansible
  executable location = /root/ansible/bin/ansible-playbook
  python version = 2.7.12 (default, Dec  4 2017, 14:50:18) [GCC 5.4.0 20160609]

TASK [ansible-haproxy : configure haproxy service] *********************************************************************************************************************************************************
fatal: [104.248.64.120]: FAILED! => {"changed": false, "msg": "AnsibleUndefinedVariable: 'ansible_eth0_2' is undefined"}
fatal: [142.93.87.190]: FAILED! => {"changed": false, "msg": "AnsibleUndefinedVariable: 'ansible_eth0_2' is undefined"}

Looking back through facts that Ansible is learning, it appears the anchor IP is exposed as ansible_eth0_1 versus ansible_eth0_2. Instead of needing to add a new task, should just need a template update.

[jonschwenn]

Can you test on an older Droplet if you happen to have one to see if it reports ansible_eth0_2? I'm curious if something changed with the image/deployments on our side.

[ful09003]

Was able to test on Debian 9.3, still reported as eth0_1. I'll make a PR to just update the jinja template.

[cmndrsp0ck]

adjusted defaults/main.yml and templates/haproxy.cfg.j2 7bf7820

passed test. closing