I try to configure advanced resource with meta and I get error "'dict object' has no attribute 'name'".
Probably the cause is nvpair.name in loop_control.labal. Should be nvpair.key.
https://github.com/devgateway/ansible-role-pacemaker/blob/master/tasks/advanced-resource.yml#L22

Hello
This is from your example playbook

pacemaker_resources:
- id: dns-ip
type: "ocf:heartbeat:IPaddr2"
options:
ip: 10.0.0.1
cidr_netmask: 8

Do i have to enter a virtual/floating IP in ip:?

thank you

The following error is observed in ansible 2.9.x:

fatal: [myhost]: FAILED! => {"changed": false, "msg": "The target XML source '/tmp/ansible.xwl9wmkg.xml' does not exist."}

This output is observed even if prior include_role calls had created the temp file already. Apparently ansible cleans up tempfiles faster than it used to.

I have submitted a patch in PR #16

Hi,
I've created my cluster manually. Nothing special.
Then i wanted to test ansible book for it. But when i run it i get this error

TASK: [styopa.pacemaker | Ensure the password for the cluster user] ***********
failed: [20cent] => {"failed": true}
msg: this module requires key=value arguments (['name={#', 'pacemaker_user', '#}', 'password={#', 'pacemaker_password', '|', 'password_hash(sha512,', 'ansible_hostname)', '#}'])
failed: [10cent] => {"failed": true}
msg: this module requires key=value arguments (['name={#', 'pacemaker_user', '#}', 'password={#', 'pacemaker_password', '|', 'password_hash(sha512,', 'ansible_hostname)', '#}'])

FATAL: all hosts have already failed -- aborting

And my playbook looks like this

• hosts: mycluster
  roles:

  • styopa.pacemaker
    vars:
    pacemaker_ansible_group: mycluster
    pacemaker_password: Support1209
    pacemaker_cluster_name: 30cent
    pacemaker_properties:
    stonith_enabled: "true"
    pacemaker_resources:

  - id: dns-ip
    type: "ocf:heartbeat:IPaddr2"
    options:
      ip: 16.53.125.50
      cidr_netmask: 8
    op:
      - action: monitor
        options:
          interval: 5s
  

Here are the variables

root@mybook:/etc/ansible/roles/styopa.pacemaker/vars# cat main.yml

pacemaker_package: pcs
pacemaker_user: hacluster

can you please explain in your playbook the - id: dns-ip and dns-server ?
are these just hostnames reflecting your set up ?

hi
i got it go past the previous error but now i am stuck here
TASK [styopa.pacemaker : Authenticate all nodes] *******************************
task path: /home/dimtheo/my-roles/styopa.pacemaker/tasks/main.yml:18
fatal: [debian3]: FAILED! => {"failed": true, "msg": "'cloud' is undefined"}

in vars/main i have these:

pacemaker_package: pacemaker
pacemaker_user: hacluster
pacemaker_ansible_group: cloud

i have defined "cloud". Any ideas why it complains ?
thank you

Hi! If, in a playbook, I first say tasks_from: group, it works perfectly. However, if I, after that, in another task, say tasks_from: constraint, it fails, e.g.: "The target XML source '/tmp/ansible.Vme64f.xml' does not exist." The 'post' task included in group.yml removes the temporary config file, whereas constraint.yml doesn't include the pre.yml (or post.yml) files. And if I run the 'constraint' task first, I get the error that 'pcmk_config' is undefined, which it is, since it never gets defined. Suggested solution: add include_tasks: pre.yml and include_tasks: post.yml to the constraint.yml file, as it is in group.yml, resource.yml and advanced_resource.yml.

It looks to me that there is nothing in the code https://github.com/styopa/ansible-pacemaker/blob/1ed86d75da60828ef1758f0b44c08cc4839f9283/tasks/main.yml#L50 to prevent from trying to create already defined pacemaker_resources on the cluster.

Take the following 2-node scenario:

1. node1 defines pacemaker_resources
2. run ansible again on node1 - pacemaker_resources will get defined again?
3. node1 fails, and node2 becomes the only nodes in the cluster
4. node1 is reinstalled from scratch - pacemaker_resources are defined again unaware of node2 being present?

hello
i get this error and the play stops

TASK [styopa.pacemaker : Install Pacemaker Configuration System package] *******
task path: /home/dimtheo/my-roles/styopa.pacemaker/tasks/main.yml:6
fatal: [debian3]: FAILED! => {"failed": true, "msg": "'pacemaker' is undefined"}
fatal: [debian4]: FAILED! => {"failed": true, "msg": "'pacemaker' is undefined"}

can you help?

Hi
i have finnaly ported your role to Debian 9 . Everything works except the part that you have to create corosync-keygen , i did that manually. I still need to figure out how to do it with ansible.

For debian8 the role fails miserably . You have to use jessie backports but even then packages don;t install for some reason.

I can share the edited role with you if you are interested

the other 2 issues still remain open ..

?

Hi,

Thank you so much for the Anisble role. It's really helpful.

I'm trying to setup Pacemaker cluster on Debian 10 and with little tweaks it's working fine.

We would like to add 2 Virtual IPs and a service for squid. I followed the taks_from: group but post tasks - Verify cluster configuration is failing with CIB did not pass DTD/schema validation.

We trying to setup something like http://wiki.stocksy.co.uk/wiki/High_Availability_IP_Addresses_in_Debian

Could you please guide how to configure our requirement please?

I am getting the following error when running the play book with the following included.
_pacemaker_private_interface: |
{% for interface in ansible_interfaces %}
{% if 'docker' in interface or 'lo' in interface %}{% continue %}{% endif %}
{% set int = 'ansible%s' | format(interface) %}
{% if _int in hostvars[inventory_hostname] and 'ipv4' in hostvars[inventory_hostname][_int] and hostvars[inventory_hostname][_int]['ipv4']['address'] is defined %}
{% if hostvars[inventory_hostname][_int]['ipv4']['address'] | ipaddr('private') %}
{{ interface|trim}}
{% break %}{% endif %}
{% endif %}
{% endfor %}

pacemaker_private_interface: "{{ _pacemaker_private_interface | trim }}"
pacemaker_corosync_ring_interface: "{{ pacemaker_private_interface }}"

here is the error;
TASK [weldpua2008.pacemaker : Creates corosync config] **************************************************
fatal: [intblade03]: FAILED! => {"changed": false, "msg": "AnsibleError: {{ pacemaker_private_interface }}: {{ _pacemaker_private_interface | trim }}: {% for interface in ansible_interfaces %}\n{% if 'docker' in interface or 'lo' in interface %}{% continue %}{% endif %}\n{% set int = 'ansible%s' | format(interface) %}\n{% if _int in hostvars[inventory_hostname] and 'ipv4' in hostvars[inventory_hostname][_int] and hostvars[inventory_hostname][_int]['ipv4']['address'] is defined %}\n {% if hostvars[inventory_hostname][_int]['ipv4']['address'] | ipaddr('private') %}\n {{ interface|trim}}\n {% break %}{% endif %}\n{% endif %}\n{% endfor %}\n: template error while templating string: Encountered unknown tag 'continue'. Jinja was looking for the following tags: 'elif' or 'else' or 'endif'. The innermost block that needs to be closed is 'if'.. String: {% for interface in ansible_interfaces %}\n{% if 'docker' in interface or 'lo' in interface %}{% continue %}{% endif %}\n{% set int = 'ansible%s' | format(interface) %}\n{% if _int in hostvars[inventory_hostname] and 'ipv4' in hostvars[inventory_hostname][_int] and hostvars[inventory_hostname][_int]['ipv4']['address'] is defined %}\n {% if hostvars[inventory_hostname][_int]['ipv4']['address'] | ipaddr('private') %}\n {{ interface|trim}}\n {% break %}{% endif %}\n{% endif %}\n{% endfor %}\n"}
fatal: [intblade04]: FAILED! => {"changed": false, "msg": "AnsibleError: {{ pacemaker_private_interface }}: {{ _pacemaker_private_interface | trim }}: {% for interface in ansible_interfaces %}\n{% if 'docker' in interface or 'lo' in interface %}{% continue %}{% endif %}\n{% set int = 'ansible%s' | format(interface) %}\n{% if _int in hostvars[inventory_hostname] and 'ipv4' in hostvars[inventory_hostname][_int] and hostvars[inventory_hostname][_int]['ipv4']['address'] is defined %}\n {% if hostvars[inventory_hostname][_int]['ipv4']['address'] | ipaddr('private') %}\n {{ interface|trim}}\n {% break %}{% endif %}\n{% endif %}\n{% endfor %}\n: template error while templating string: Encountered unknown tag 'continue'. Jinja was looking for the following tags: 'elif' or 'else' or 'endif'. The innermost block that needs to be closed is 'if'.. String: {% for interface in ansible_interfaces %}\n{% if 'docker' in interface or 'lo' in interface %}{% continue %}{% endif %}\n{% set int = 'ansible%s' | format(interface) %}\n{% if _int in hostvars[inventory_hostname] and 'ipv4' in hostvars[inventory_hostname][_int] and hostvars[inventory_hostname][_int]['ipv4']['address'] is defined %}\n {% if hostvars[inventory_hostname][_int]['ipv4']['address'] | ipaddr('private') %}\n {{ interface|trim}}\n {% break %}{% endif %}\n{% endif %}\n{% endfor %}\n"}

hi
i get stuck here

• name: Authenticate all nodes
  command: >
  pcs cluster auth
  {% for host in groups[pacemaker_ansible_group] %}
  {{ hostvars[host]['ansible_hostname'] }}
  {% endfor %}
  -u {{ pacemaker_user }} -p {{ pacemaker_password }}
  run_once: true
  args:
  creates: /var/lib/pcsd/tokens

it craps out with a message saying it can not create the above dir
can you give some help please?

fatal: [debian3]: FAILED! => {"changed": false, "cmd": "pcs cluster auth debian3 debian4 -u hacluster -p secret", "failed": true, "invocation": {"module_args": {"_raw_params": "pcs cluster auth debian3 debian4 -u hacluster -p secret", "_uses_shell": false, "chdir": null, "creates": "/var/lib/pcsd/tokens", "executable": null, "removes": null, "warn": true}, "module_name": "command"}, "msg": "[Errno 2] No such file or directory", "rc": 2}