b171ccbe07
In 641858783778ee14b627a55df48f81ad40c94555 we removed the extra vars from several roles. Well it's actually important in just one, and only sometimes. Anyways if and only if the dns server gets set incorrectly on nodes (happens when the lab is having trouble) collectd adds a DNS server, 8.8.8.8 by default which means we go from "it might work eventually" to "guaranteed to never work" once CollectD runs without the dns_server var changed from default. This went unnoticed because days can pass between issues like this. Anyways alls the builds are failing now on it, so need to readd it at least here. Change-Id: I5a32562e19415d29764c9b22aec95fa3dc95ad85
20 lines
669 B
YAML
20 lines
669 B
YAML
---
|
|
#role to install CollectD on all nodes
|
|
|
|
- name: Install CollectD
|
|
shell:
|
|
"cd {{ ansible_env.HOME }}/browbeat/ansible; \
|
|
ansible-playbook -i hosts \
|
|
--extra-vars collectd_compute=true \
|
|
--extra-vars graphite_host={{ graphite_host_template }} \
|
|
--extra-vars graphite_prefix={{ graphite_prefix_template }} \
|
|
--extra-vars dns_server={{ dns_server }} \
|
|
install/collectd-openstack.yml \
|
|
> {{ ansible_env.HOME }}/browbeat/results/collecd_install.log"
|
|
register: collectd_install
|
|
until: collectd_install.rc == 0
|
|
retries: 2
|
|
delay: 60
|
|
environment:
|
|
ANSIBLE_SSH_ARGS: "-F {{ ansible_env.HOME }}/browbeat/ansible/ssh-config"
|