I want to create playbook that going on the entire inventory file and divide the servers to 2 groups: Windows and Linux (add-host going only on the first server on the inv)
I tried this code:
- name: Linux Group
add_host:
name: "{{ item }}"
group: LinuxGroup
when: hostvars[" {{ item }} "]['ansible_system'] == 'Linux'
with_items: "{{ ansible_play_hosts_all }} "
This code suppose to create the linux group
and I tried with other conditions of when but non was a succsess. I would like to get your help please.
PS:
I changed the code to this:
tasks:
- name: Create linux group
add_host:
name: "{{ item }}"
group: LinuxGroup
when: hostvars[item].ansible_system == 'Linux'
with_items: "{{ ansible_play_hosts_all }} "
ignore_errors: yes
- name: ping to Linux
ping:
with_items: LinuxGroup
and when I run the code the windows servers are skipped in the "Create linux group" task, but I used the module debug to print to group's items and there are the windows servers.
The fact ansible_system is a key under the host, not a list of keys. Also, conditional statements should not include jinja2 templating delimiters such as {{ }} or {% %}, use hostvars[item] instead of hostvars[ "{{ item }}" ].
Make sure gather_facts is set to true.
Note that the add_host module does not update the inventory file but updates the in-memory inventory. I have included a debug task to print the groups from the in-memory inventory of ansible.
gather_facts: true
tasks:
- name: Linux Group
add_host:
name: "{{ item }}"
group: LinuxGroup
when: hostvars[item].ansible_system == 'Linux'
with_items: "{{ ansible_play_hosts_all }}"
- debug: msg="{{ groups }}"
Related
I have an inventory that i need to add a conditional to.
My code:
- name: Create memory inventory
add_host:
name: "{{ item.0.key }}"
group: target_hosts
with_nested:
- "{{ lookup ('dict', hosts, wantlist=True) }}"
But I want something like:
- name: Create memory inventory
add_host:
name: "{{ item.0.key }}"
{% if item.0.value.OS_Choice[:3] == 'win' %}
group:
- target_hosts
- win
{% else %}
group:
- target_hosts
- linux
{% endif %}
with_nested:
- "{{ lookup ('dict', hosts, wantlist=True) }}"
With this configuration, Ansible errors:
The offending line appears to be:
{% if item.0.value.OS_Choice[:3] == 'win' %}
^ here
Any ideas on how I can implement this conditional?
You're mixing up Jinja2 with YAML. Here you go:
- name: Create memory inventory when win
add_host:
name: "{{ item.0.key }}"
with_nested:
- "{{ lookup ('dict', hosts, wantlist=True) }}"
when: item.0.value.OS_Choice[:3] == 'win'
vars:
group:
- target_hosts
- win
- name: Create memory inventory when not win
add_host:
name: "{{ item.0.key }}"
with_nested:
- "{{ lookup ('dict', hosts, wantlist=True) }}"
when: item.0.value.OS_Choice[:3] != 'win'
vars:
group:
- target_hosts
- linux
However, Ansible gather facts about the OS already. Perhaps you want to use those, instead of configuring something like this yourself.
Building up on Kevin's answer (and fixing some wrongly placed parameters)
You should definitely do this differently, like creating dynamic groups based on detected OS in facts. See:
the group_by module
the ansible_distribution* facts that you can explore as an example with
ansible localhost -m setup -a filter="ansible_distribution*"
Meanwhile, with your current logic, you can still do this in a single task:
- name: Create memory inventory
vars:
additional_group: >-
{{ (item.0.value.OS_Choice[:3] == 'win') | ternary('win', 'linux') }}
add_host:
name: "{{ item.0.key }}"
groups:
- target_hosts
- "{{ additional_group }}"
with_nested:
- "{{ lookup ('dict', hosts, wantlist=True) }}"
I have below inventory file:
[web]
10.0.1.0
[db]
10.0.3.0
[frontend]
10.0.5.0
[X_all:children]
web
db
frontend
Now in my playbook: I'm trying to print the hostname under X_all group in the name line.
- name: "Copying the output of Registry run to local machine from {{ hostvars[item]['inventory_hostname'] }}"
become: true
become_user: "{{ login }}"
fetch:
src: /tmp/DeploymentRegistry.txt
dest: /tmp/DeploymentRegistry-{{ inventory_hostname }}.txt
flat: yes
with_items:
- "{{ groups['X_all'] }}"
Is it even possible to add the hostname in the name line?
I tried replacing {{ hostvars[item]['inventory_hostname'] }} with inventory_hostname but it's also not working.
Regarding your requirement
I'm trying to print the hostname under test group in the name line.
take note that the group has more than one member and therefore more than one hostname, but a list of hostnames.
- name: Show hosts {{ groups['test'] }}
delegate_to: localhost
debug:
msg: "{{ item }}" # will show the group member hostname(s)
with_items:
- "{{ groups['test'] }}"
Is it even possible to add the hostname in the name line? I tried ... with inventory_hostname but it's also not working.
Yes, it is.
- name: Show host {{ inventory_hostname }}
debug:
msg: "{{ item }}"
with_items:
- "{{ groups['test'] }}"
But it will print the first hostname from list only.
I'm trying to pop VM instances, put them into different host groups (say webservers and devops/admin machines) and install what is needed on them in one single playbook.
I don't know what IP addresses, for instance, GCP will give these instances, and so i am trying to capture them in a variable for use later on in the playbook. I can capture them fine by using "register" but using them is proving tricky. For instance if I do.
- name: création des adresses statiques
gcp_compute_address:
name: "{{ item }}"
state: present
region: "{{ region }}"
project: "{{ gcp_project }}"
auth_kind: "{{ gcp_cred_kind }}"
service_account_file: "{{ gcp_cred_file }}"
scopes:
- https://www.googleapis.com/auth/compute
loop:
- adresse-1
- adresse-2
- adresse-3
- adresse-4
- adresse-5
register: address
The best way i have figured out to use these variables later on is:
network_interfaces:
- network: "{{ network.name }}"
access_configs:
- name: 'External NAT'
type: 'ONE_TO_ONE_NAT'
nat_ip:
- "{{ address.results[0].address }}"
- "{{ address.results[1].address }}"
- "{{ address.results[2].address }}"
Which fails miserably.
Please help ? How can I use the range of addresses I have created ?
I am going nuts over this
It is possible to add_host to the group webservers and proceed with the next play
- add_host:
name: "{{ item }}"
groups: webservers
loop: "{{ address.results|json_query('[*].address') }}"
- debug:
msg: "{{ groups['webservers'] }}"
- hosts: webservers
tasks:
- name: Configure cluster
...
The tasks below split the hosts into two groups
- set_fact:
my_hosts: "{{ address.results|json_query('[*].address') }}"
- add_host:
name: "{{ item }}"
groups: webservers1
loop: "{{ my_hosts[0:(my_hosts|length / 2)|int] }}"
- add_host:
name: "{{ item }}"
groups: webservers2
loop: "{{ my_hosts[(my_hosts|length / 2)|int:my_hosts|length] }}"
- debug:
msg: "{{ groups['webservers1'] }}"
- debug:
msg: "{{ groups['webservers2'] }}"
There is also GCE Dynamic Inventory and other 100+ gcp modules. You might want to start with Google Cloud Platform Guide.
FWIW, Ansible 2 Cloud Automation Cookbook covers leading providers incl. GCP.
I have an assignment to provision {{ create new hosts }} according to a given
invonteroy file,
example
[SET1]
1.1.1.1 new_ip=1.1.1.1
[SET2]
1.1.1.2 new_ip=1.1.1.2
[SET3]
1.1.1.3 new_ip=1.1.1.3
Here is the problem. The inventory file contains set of ip addresses that do not exists yet.
What I want to do
- connect to vsphere host(5.5)
- create new machines based on amount of hosts in the inventory
- change the machines ip address to match the one in the given inventory
modules like
vmware_shell
vmware_guest
requires administrative permissions which I don't have.
The only way to hack around this issue is to try and create an in memory or temporary inventory during provisioning task.
Here is my role:
- vsphere_guest:
vcenter_hostname: {{ vcenter }}
validate_certs: no
username: "{{ vsphere.username }}"
password: "{{ vsphere.password }}"
guest: "{{ item }}"
from_template: yes
template_src: Centos_base
resource_pool: "/Resources"
esxi:
datacenter: Remote
hostname: {{ esx_host }}
with_items:
- "{{ play_hosts }}"
- name: wait for power on to finish
pause: seconds=5
- name: get facts from vm
vsphere_guest:
validate_certs: False
vcenter_hostname: {{ vcenter }}
username: "{{ vsphere.username }}"
password: "{{ vsphere.password }}"
guest: "{{ item }}"
vmware_guest_facts: yes
with_items:
- "{{ play_hosts }}"
register: vm_facts
until: vm_facts.ansible_facts.hw_eth0.ipaddresses[0] is defined
retries: 10
delay: 10
- name: grab new ipaddresses
set_fact: myIp="{{ inventory_hostname }}"
- name: Add production VMs to inventory
add_host: hostname={{item.ansible_facts.hw_eth0.ipaddresses[0] }} groups=temp
new_ip={{ myIp }}
with_items: "{{ vm_facts.results }}"
ignore_errors: yes
my question:
is there any way to create a temp inventory with multiple groups and mutilple variables that matches the given inventory!?
update
I hacked a way around it:
- name: Add production VMs to inventory
add_host: hostname={{item.ansible_facts.hw_eth0.ipaddresses[0] }} groups=temp
new_ip="{{ item.ansible_facts.hw_name }}"
with_items: "{{ vm_facts.results }}"
ignore_errors: yes
this works since the machine name is equal to {{ play_host }}
Based on this question
Ansible recursive checks in playbooks
I have another one.
We need to go through this structure
Zone spec https://gist.github.com/git001/9230f041aaa34d22ec82eb17d444550c
Now I can adress the hostnames via the array index but can I also iterate over the array "hosts"?
playbook
--
- hosts: all
gather_facts: no
vars_files:
- "../doc/application-zone-spec.yml"
roles:
- { role: ingress_add, customers: "{{ application_zone_spec }}" }
role
- name: Print ingress hostnames
debug: msg="{{ item.hosts.0.hostname }} {{ item.hosts.1.hostname }}"
with_items: "{{ customers.ingress }}"
We use.
ansible-playbook --version
ansible-playbook 2.1.0.0
config file = /etc/ansible/ansible.cfg
configured module search path = Default w/o overrides
Use with_subelements:
- name: Print ingress hostnames
debug: msg="{{ item.0.type }} {{ item.1.hostname }}"
with_subelements:
- "{{ customers.ingress }}"
- "hosts"
There is quite a bit of examples for different loops in the Loops section of the documentation.