ansible loop with failed_when on register variable list output - ansible

Question

Team,
I have below task working fine when i only have one node or only one item but I need to modify it work for all items returned in the list that is stored in register variable.
- name: "Fetch all CPU nodes from clusters using K8s beta.kubernetes.io/instance-type"
k8s_info:
kind: Node
label_selectors:
- "beta.kubernetes.io/instance-type={{ kube_cpu_node_class }}"
verify_ssl: no
register: cpu_class_list
failed_when: not cpu_class_list.resources
How can i do this for all nodes in cpu_class_list variable with loop or with_items?
proposed solution but does not work
- name: "Fetch all CPU nodes from clusters using K8s beta.kubernetes.io/instance-type"
k8s_info:
kind: Node
label_selectors:
- "beta.kubernetes.io/instance-type={{ kube_cpu_node_class }}"
verify_ssl: no
register: cpu_class_list
failed_when: not {{ item }}
with_items: cpu_class_list.resources
sample output with two nodes is below
services-pre-install-checks : debug]
ok: [localhost] => {
"cpu_class_list": {
"changed": false,
"deprecations": [
{
"msg": "The 'k8s_facts' module has been renamed to 'k8s_info'",
"version": "2.13"
}
],
"failed": false,
"failed_when_result": false,
"resources": [
{
"apiVersion": "v1",
"kind": "Node",
"metadata": {
"annotations": {
"volumes.kubernetes.io/controller-managed-attach-detach": "true"
},
"creationTimestamp": "2019-07-16T00:23:27Z",
"labels": {
"nodeType": "cpu"
},
"name": "node1",
"nodeInfo": {
"architecture": "amd64",
}
}
},
{
{
"apiVersion": "v1",
"kind": "Node",
"metadata": {
"annotations": {
"volumes.kubernetes.io/controller-managed-attach-detach": "true"
},
"creationTimestamp": "2019-07-16T00:23:27Z",
"labels": {
"nodeType": "cpu"
},
"name": "node2",
"nodeInfo": {
"architecture": "amd64",
}
}
}
]
}
}
Proposed Solution:
- name: "Fetch all CPU nodes from clusters using K8s beta.kubernetes.io/instance-type"
k8s_facts:
kind: Node
label_selectors:
- "beta.kubernetes.io/instance-type={{ kube_cpu_node_class }}"
verify_ssl: no
register: cpu_class_list
failed_when: not cpu_class_list.resources
#above to fail when none of the nodes has label, that is resources list is empty.
#below to fail when any of the nodes has no label
- debug:
msg: "{{ item.metadata.labels.nodeType }}"
loop: "{{ cpu_class_list.resources }}"
loop_control:
label: "{{ item.metadata.name }}"
failed_when: not item.metadata.labels.nodeType

Answer 1

I would propose to split it up into two distinct tasks. First, register the variable, then use the fail ansible module (docs) to inspect the var and fail if a condition is met.
See this snippet for outlining the logic:
- hosts: localhost
vars:
test: # test array
- fail: false
- fail: false
- fail: true
- fail: false
tasks:
- name: iterate and fail
fail:
msg: "Fail as requested"
with_items: "{{ test }}"
when: item.fail
Running this, will output the following:
$ ansible-playbook failing.yml
PLAY [localhost] ***********************
TASK [Gathering Facts] *************************
ok: [localhost]
TASK [iterate and fail] **************************
skipping: [localhost] => (item={u'fail': False})
skipping: [localhost] => (item={u'fail': False})
failed: [localhost] (item={u'fail': True}) => {"ansible_loop_var": "item", "changed": false, "item": {"fail": true}, "msg": "Failed as requested"}
skipping: [localhost] => (item={u'fail': False})
PLAY RECAP *****************
localhost : ok=1 changed=0 unreachable=0 failed=1 skipped=0 rescued=0 ignored=0
Hope this helps!

Answer 2

#to fail when none of the nodes has label, that is resources list is empty.
- name: "Fetch all CPU nodes from clusters using K8s beta.kubernetes.io/instance-type"
k8s_facts:
kind: Node
label_selectors:
- "beta.kubernetes.io/instance-type={{ kube_cpu_node_class }}"
verify_ssl: no
register: cpu_class_list
failed_when: not cpu_class_list.resources
#below to fail when any of the nodes has no label
- debug:
msg: "{{ item.metadata.labels.nodeType }}"
loop: "{{ cpu_class_list.resources }}"
loop_control:
label: "{{ item.metadata.name }}"
failed_when: not item.metadata.labels.nodeType