I have the following with_nested:
- name: Create solr schema for solr_cores
uri:
url: http://{{ cassandra_cluster_ips.split(',') | random }}:8983/solr/admin/cores?action={{ solr_core_action }}&core={{ item[1] }}.{{ item[0] }}
timeout: "{{ solr_create_timeout }}"
sudo: True
with_nested:
- "{{ solr_cores }}"
- "{{ client_names }}"
I want to change to my extra vars from:
#solr_cores: ['dom_chunk_meta', 'dom', 'tra_chunk_meta','tra','dom_difference_results','me_output']
#solr_core_action: "reload"
to:
solr_cores: [{‘dom_difference_results’:‘create’}, {‘dom’:‘reload’},
{‘tra’:‘reload’}, {‘me_output’:‘reload’}]
I looked at subelements, but don’t know how to pass it as a simple dictionary list so that I can set them into uri up to access it.
I am not 100% sure what you are trying to do but with_dict might be the one to use. If this doesn't work for you, comment and I will adjust the answer.
In your vars/inventory:
solr:
- name: core1
create_timeout: 30
action: reload
- name: core2
action: create
create_timeout: 60
In your playbook:
- name: Create solr schema for solr_cores
uri:
url: http://{{ cassandra_cluster_ips.split(',') | random }}:8983/solr/admin/cores?action={{ item.value.core_action }}&core={{ item.value.name }}.{{ item.value.other }}
timeout: "{{ item.value.create_timeout }}"
sudo: True
with_dict: "{{ solr }}"
Related
Below is the condition
- name: Find the image
slurp:
src: "{{ IMAGE }}"
register: slurp_results
- name: Upload image
shell: |
skopeo copy -docker-archive:{{ item }}.tar docker://{{ URL }}/TESTIMAGE
with_items: "{{ (slurp_results.content|b64decode).splitlines() }}"
The above code works.
But I would need "TESTIMAGE" also to be replaced as {{ item }} like below.
skopeo copy -docker-archive:{{ item }}.tar docker://{{ URL }}/{{ item }}
How to define 2 with_items in a single shell task with 2 different slurp results
I believe you can by using the subelements module. Here is a link. Try going by this example:
- name: Setup MySQL users, given the mysql hosts and privs subkey lists
mysql_user:
name: "{{ item.0.name }}"
password: "{{ item.0.mysql.password }}"
host: "{{ item.1 }}"
priv: "{{ item.0.mysql.privs | join('/') }}"
with_subelements:
- "{{ users }}"
- mysql.hosts
Users is referred to as item.0 and hosts as item.1 and so on.
I am looking to loop through a list of variables. I have it looping through the of variables using with_items, however the catch is there is a list within that variables list that needs to have a different subset / number of variables that i need to iterate through as well.
I have tried different filters to include with_nested, with_subelements, and with_items. I know that they are moving towards loops as the primary driver moving forward so any solution ideally would leverage the ansible path moving forward. I am looking at having an "inner" loop or an external task that will iterate through the vlans_list and input that data as its to that point.
group Variables
vnic_templates:
- name: vNIC-A
fabric: A
mac_pool: testmac1
mtu: 1500
org_dn: org-root
redundancy_type: none
state: present
template_type: initial-template
vlans_list: ### THE PROBLEM CHILD
- name: vlan2
native: 'no'
state: present
- name: vlan3
native: 'no'
state: present
The actual task - i have issues when i have to input multiple vlans. The vnic template will have a 1 to one relationship however the vlans_list could be 1 vnic_template to many vlans.
ucs_vnic_template:
hostname: "{{ ucs_manager_hostname }}"
username: "{{ ucs_manager_username }}"
password: "{{ ucs_manager_password }}"
name: "{{ item.name }}"
fabric: "{{ item.fabric }}"
mac_pool: "{{ item.mac_pool }}"
mtu: "{{ item.mtu }}"
org_dn: "{{ item.org_dn }}"
redundancy_type: "{{ item.redundancy_type }}"
state: "{{ item.state }}"
template_type: "{{ item.template_type }}"
vlans_list:
- name: "{{ item.1.name }}"
native: "{{ item.1.native }}"
state: "{{ item.1.present }}"
# loop: "{{ vnic_templates | subelements('vlans_list') }}"
with_items:
- "{{ vnic_templates }}"
I am starting down the road of adding an include vlan_list.yml outside of this task but no familiar with out to do that.
Actual results are
The task includes an option with an undefined variable. The error was: 'item' is undefined\n\n
I need the create a single vnic template with multiple vlans defined in that list.
Another engineer i work with was able to solve the question. By the way the variables are laid out we were able to easily just change the code
Change this:
vlans_list:
- name: "{{ item.1.name }}"
native: "{{ item.1.native }}"
state: "{{ item.1.present }}"
To this:
vlans_list: "{{ item.vlans_list }}"
Full Code listed below.
- name: Add vNIC Templates
ucs_vnic_template:
hostname: "{{ ucs_manager_hostname }}"
username: "{{ ucs_manager_username }}"
password: "{{ ucs_manager_password }}"
name: "{{ item.name }}"
fabric: "{{ item.fabric }}"
mac_pool: "{{ item.mac_pool }}"
mtu: "{{ item.mtu }}"
org_dn: "{{ item.org_dn }}"
redundancy_type: "{{ item.redundancy_type }}"
state: "{{ item.state }}"
template_type: "{{ item.template_type }}"
vlans_list: "{{ item.vlans_list }}"
with_items:
- "{{ vnic_templates }}"
I have the following dict:
endpoint:
esxi_hostname: servername.domain.com
I'm trying to use it as an option via jinja2 for the vmware_guest but have been unsuccessful. The reason I'm trying to do it this way is because the dict is dynamic...it can either be cluster: clustername or esxi_hostname: hostname, both mutually exclusive in the vmware_guest module.
Here is how I'm presenting it to the module:
- name: Create VM pysphere
vmware_guest:
hostname: "{{ vcenter_hostname }}"
username: "{{ username }}"
password: "{{ password }}"
validate_certs: no
datacenter: "{{ ansible_host_datacenter }}"
folder: "/DCC/{{ ansible_host_datacenter }}/vm"
"{{ endpoint }}"
name: "{{ guest }}"
state: present
guest_id: "{{ osid }}"
disk: "{{ disks }}"
networks: "{{ niclist }}"
hardware:
memory_mb: "{{ memory_gb|int * 1024 }}"
num_cpus: "{{ num_cpus|int }}"
scsi: "{{ scsi }}"
customvalues: "{{ customvalues }}"
cdrom:
type: client
delegate_to: localhost
And here is the error I'm getting when including the tasks file:
TASK [Preparation : Include VM tasks] *********************************************************************************************************************************************************************************
fatal: [10.10.10.10]: FAILED! => {"reason": "Syntax Error while loading YAML.
The error appears to have been in '/data01/home/hit/tools/ansible/playbooks/roles/Preparation/tasks/prepareVM.yml': line 36, column 4, but may
be elsewhere in the file depending on the exact syntax problem.
The offending line appears to be:
"{{ endpoint }}"
hostname: "{{ vcenter_hostname }}"
^ here
We could be wrong, but this one looks like it might be an issue with
missing quotes. Always quote template expression brackets when they
start a value. For instance:
with_items:
- {{ foo }}
Should be written as:
with_items:
- "{{ foo }}"
exception type: <class 'yaml.parser.ParserError'>
exception: while parsing a block mapping
in "<unicode string>", line 33, column 3
did not find expected key
in "<unicode string>", line 36, column 4"}
So in summary, I'm not sure how to format this or if it is even possible.
The post from techraf sums up your problem, but for a possible solution, in the docs, especially regarding Jinja filters, there is the following bit:
Omitting Parameters
As of Ansible 1.8, it is possible to use the default filter to omit
module parameters using the special omit variable:
- name: touch files with an optional mode
file: dest={{item.path}} state=touch mode={{item.mode|default(omit)}} > with_items:
- path: /tmp/foo
- path: /tmp/bar
- path: /tmp/baz
mode: "0444"
For the first two files in the list, the default mode will be
determined by the umask of the system as the mode= parameter will not
be sent to the file module while the final file will receive the
mode=0444 option.
So it looks like what should be tried is:
esxi_hostname: "{{ endpoint.esxi_hostname | default(omit) }}"
# however you want the alternative cluster settings done.
# I dont know this module.
cluster: "{{ cluster | default(omit) }}"
This is obviously reliant on the vars to only have one choice set.
There is no way you could ever use the syntax you tried in the question, because firstly and foremostly Ansible requires a valid YAML file.
The closest workaround would be to use a YAML anchor/alias although it would work only with literals:
# ...
vars:
endpoint: &endpoint
esxi_hostname: servername.domain.com
tasks:
- name: Create VM pysphere
vmware_guest:
hostname: "{{ vcenter_hostname }}"
username: "{{ username }}"
password: "{{ password }}"
validate_certs: no
datacenter: "{{ ansible_host_datacenter }}"
folder: "/DCC/{{ ansible_host_datacenter }}/vm"
<<: *endpoint
name: "{{ guest }}"
state: present
guest_id: "{{ osid }}"
disk: "{{ disks }}"
networks: "{{ niclist }}"
hardware:
memory_mb: "{{ memory_gb|int * 1024 }}"
num_cpus: "{{ num_cpus|int }}"
scsi: "{{ scsi }}"
customvalues: "{{ customvalues }}"
cdrom:
type: client
delegate_to: localhost
I am currently working on a host where i have installed ansible. I have created 2 application accounts with groups with nologin and within that groups i want to add users, so that every department has their own ansible directory.
My vars look like below:
---
- hosts: localhost
become: yes
vars:
ansible_groupuser:
- name: "ansible-dictators"
ansible_groupuser_uid: "3000"
ansible_users:
- idia
- josefs
- donaldt
- kimjongu
- name: "ansible-druglords"
ansible_groupuser_uid: "3001"
ansible_users:
- pabloe
- javierg
- frankl
- rossu
Now i have 2 plays. 1 to create the Groupuser:
# This creates the groupuser
- name: Play 1 Create central ansible user and group per department
user:
name: "{{ item.name }}"
shell: "/sbin/nologin"
home: "/home/{{ item.name }}"
comment: "{{ item.name }} Group Account"
uid: "{{ item.ansible_groupuser_uid }}"
append: "yes"
with_items:
- "{{ansible_groupuser}}"
And 1 to create the "normal" users:
- name: Play 2 Create users
user:
name: "{{ item.1 }}"
shell: "/bin/bash"
home: "/home/{{ item.1 }}"
comment: "{{ item.1 }}"
groups: "{{ item.0.name }}"
append: "yes"
with_subelements:
- "{{ ansible_groupuser }}"
- ansible_users
If i run this play it creates the groupuser ansible-dictators on 3000 and ansible-druglords on 3001. idia gets 3002, josefs gets 3003 etc. It gets kinda messy, when i want to add a 3th groupuser like ansible-rockstars, it starts counting at the first available uid, 3010. What i want is to place the groupusers and the common users in 2 different ranges (2000 and 3000 for example)
When i do a with_together on the first play, like below, it works:
- name: Play1 Create central ansible user and group per department
user:
name: "{{ item.0.name }}"
shell: "/sbin/nologin"
home: "/home/{{ item.0.name }}"
comment: "{{ item.0.name }} Group Account"
uid: "{{ item.1 }}"
append: "yes"
with_together:
- "{{ansible_groupuser}}"
- "{{ range(3000,3020)|list }}"
when: item.0 != None
But when i do a with_together on the second play, it doesnt work:
- name: Create users
user:
name: "{{ item.1 }}"
shell: "/bin/bash"
home: "/home/{{ item.1 }}"
comment: "{{ item.1 }}"
groups: "{{ item.0.name }}"
append: "yes"
uid: "{{ item.2 }}"
with_together:
- "{{ ansible_groupuser }}"
- ansible_users
- "{{ range(2000,2020)|list }}"
Anyone got a suggestion how to make the second play work with a uid in a certain range? Or another suggestion how to get the uid's in different groups? To give the groupusers an uid in the vars is no problem. But i am expecting a lot of "common" users to add (+50) and i dont want to specify a uid for all of those users.
Hope it makes sense. Thanks in advance.
I think range(...) approach has a flaw: if you delete some user from your list in the future, IDs for subsequent entries will change and you can end up with messed file permissions on your system.
You can patch user module to support --firstuid/--lastuid arguments of the underlying adduser command, so you can set different range for uid generation.
But I'd suggest you to define "static" uids for top-level users in your vars file (from some predefined range, say: 3000..30xx) – this way you can safely add/remove top-level user/groups in the future.
And leave "common" users to get their ids automatically, so adding/deleting them will not mess your ids. If you like them to be from some specific range, you can modify system-wide /etc/adduser.conf with FIRST_UID=5000/LAST_UID=6000.
To start off I have all my variables defined in YAML
app_dir: "/mnt/{{ item.name }}"
app_dir_ext: "/mnt/{{ item.0.name }}"
workstreams:
- name: tigers
service_workstream: tigers-svc
app_sub_dir:
- inbound
- inbound/incoming
- inbound/error
- inbound/error/processed
- name: lions
service_workstream: lions-svc
app_sub_dir:
- inbound
- inbound/incoming
- inbound/error
- inbound/error/processed
You may note app_dir: "/mnt/{{ item.name }}" and app_dir_ext: "/mnt/{{ item.0.name }}" looking odd, so I originally had my variables set as below in YAML but decided to use the above mainly due to less lines in YAML when I have a large amount of workstreams.
workstreams:
- name: tigers
service_workstream: tigers-svc
app_dir: /mnt/tigers
...
I then have Ansible code to check if the directories exists, if not create them and apply permissions (!note, have taken this approach due to a ssh timeout on operation when using the file: module on a number of very big NFS mounted shares).
- name: Check workstreams app_dir
stat:
path: "{{ app_dir }}"
register: app_dir_status
with_items:
- "{{ workstreams }}"
- name: Check workstreams app_sub_dir
stat:
path: "{{ app_dir_ext }}/{{ item.1 }}/"
register: app_sub_dir_status
with_subelements:
- "{{ workstreams }}"
- app_sub_dir
- name: create workstreams app_dir
file:
path: "/mnt/{{ item.0.name }}"
state: directory
owner: "ftp"
group: "ftp"
mode: '0770'
recurse: yes
with_nested:
- '{{ workstreams }}'
- app_dir_status.results
when:
- '{{ item.1.stat.exists }} == false'
This is a little hacky but works, however I have a 2nd, 3rd, 4th path to check..etc
My question here is how to I update/refactor the above code to use <register_name>.stat.exists == false from both app_dir_status and app_sub_dir_status to control my task ?
You don't need to make nested loop! There's all required data inside app_sub_dir_status – just strip unnecessary items.
Here's simplified example:
---
- hosts: localhost
gather_facts: no
vars:
my_list:
- name: zzz1
sub:
- aaa
- ccc
- name: zzz2
sub:
- aaa
- bbb
tasks:
- stat:
path: /tmp/{{ item.0.name }}/{{ item.1 }}
register: stat_res
with_subelements:
- "{{ my_list }}"
- sub
- debug:
msg: "path to create '{{ item }}'"
with_items: "{{ stat_res.results | rejectattr('stat.exists') | map(attribute='invocation.module_args.path') | list }}"
You can iterate over stat_res.results | rejectattr('stat.exists') | list as well, but will have to construct path again as /tmp/{{ item.item.0.name }}/{{ item.item.1 }} – note double item, because first item is an element of stat_res.results which contain another item as element of your original loop for stat task.
P.S. Also I see no reason for your first task, as subdir task can detect all missing directories.