Otdfbt

Need to register results from loop (task #1) and check when command.results stderr is not empty(task #2) relatively to loop items, so I can use this condition at new tasks.

The problem is that:

when condition from task #2 doesn't work cause I can't understand how to divide command.results relatively to loop items.

Is there solution to work around?

---
- name: "CHECK IF EXIST OPENVZ CONTAINER"
 hosts: "projectname-OPENVZHOSTS"
 gather_facts: yes
 remote_user: root
 vars:
 openvz_check: "yes"
 openvz_instances:
 opwnvz_instance_1: 
 vmid: "119"
 opwnvz_instance_2: 
 vmid: "118"

 tasks:

 - name: "Ping my hosts"
 ping:


 - name: "1. Register exist containers"
 shell: echo $(vzlist -a item.value.vmid )
 register: command
 when: openvz_check == "yes"
 loop: " lookup('dict', openvz_instances) "
 ignore_errors: yes

 - name: "2. Create variable from command"
 debug: msg=" list "
 when: 
 - openvz_check == "yes" 
 - command.results | map(attribute='stderr') | list != "" 
 loop: " lookup('dict', openvz_instances) "
 ignore_errors: yes

 - debug: msg="list "

Debug of command.results | map(attribute='stderr') | list:

ok: [projectname-OPENVZHOST-S1] => 
 "msg": [
 [], 
 []
 ]

Please get advice or solution as you see. Or what I need to research.

Regards

PS: have an idea to create folder and then check if it exist or not exist, but have no idea how to relation dict to results

You want to use the length filter.

- command.results | map(attribute='stderr') | list | length != 0

Or even better, just access stderr directly.

- command.stderr != ""

Full playbook I used is here:

---

- name: test playbook
 hosts: all
 vars:
 my_list_var: []
 tasks:
 - name: debug
 debug:
 msg: "list is empty"
 when:
 - my_list_var | list | length == 0
 - name: ls command
 command: "ls -l /tmp"
 register: result
 - name: debug 2
 debug:
 msg: "no errors"
 when: result.stderr == ""