I need to have all the userid in a single variable, all separated by \n.
Code is as below.
- name: Retrieve the user id and instance
shell: ls -l {{item}} | grep -v total| awk '{print $3}'
register: find_result_userid
with_items:
- /tmp/log/logs/log1
- /tmp/log/logs/log2
- /tmp/log/logs/log3
- name: Combine all userid
set_fact:
server_names: "{{ find_result_userid.results | map(attribute='stdout_lines')|list }}"
The output is as below.
ok: [localhost] => {
"ansible_facts": {
"server_names": [
[
"root",
"root",
"root"
],
[
"root",
"root",
"root"
],
[
"root",
"root",
"root"
]
]
},
"changed": false
}
I need something like below: i.e all ids separated by a line in a single variable.
"server_names": [
[
"root",
"root",
"root",
"root",
"root",
"root",
"root",
"root",
"root"
]
Kindly advise.
flatten the lists
- set_fact:
server_names: "{{ server_names|flatten }}"
If the number of items you are iterating with is static, I would guess, that you can use the + operator to append the your results
- name: Combine all userid
set_fact:
server_names: "{{ find_result_userid.results[0].stdout_lines + find_result_userid.results[1].stdout_lines + find_result_userid.results[2].stdout_lines}}"
Otherwise if it is not static, I think Vladimir Botkas answer is better.
Combining with what Vladmir Botka suggested, to get result in a single task.
- name: Combine all userid
set_fact:
server_names: "{{ find_result_userid.results | map(attribute='stdout_lines')|list | flatten }}"
Related
given this dictionary:
passes:
sql_x:
password:
sql_y:
password:
I want to create random passwords for any key in the passes dict.
How do I loop through the keys in the dict and fill the password value with a random password?
I was able to do it with a list but I need to use a dict.
Something like this:
- name: create passwords
set_fact: "{{ item.value.password}}": "{{ lookup('password', '/dev/null', seed=inventory_hostname) }}"
loop: "{{ lookup('dict', passes) }}"
This code above does not work of course, just for clearance what I am trying to achieve.
Thanks for any hint.
you loop over dict2items
- name: "make this working"
hosts: localhost
vars:
passes:
sql_x:
password:
sql_y:
password:
tasks:
- name: Debug
set_fact:
passes: "{{ passes | combine ({item.key: {'password': password}}) }}"
loop: "{{ passes | dict2items }}"
vars:
password: "{{ lookup('password', '/dev/null') }}"
- name: display
debug:
msg: "{{ passes }}"
result:
ok: [localhost] => {
"msg": {
"sql_x": {
"password": "kcOqz_mbIiiT0Wo_2Qox"
},
"sql_y": {
"password": "TMN_nKbnAEIzI5w-8Of."
}
}
}
I have a below value in a variable allcsv
"msg": [
{
"added_bundle_images": [
"locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.8.3-10"
],
"index_image": "locuz.qe.pnq.local/eng-build/iib:297699",
"locuz_version": "v4.8"
},
{
"added_bundle_images": [
"locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.10.0-57"
],
"index_image": "locuz.qe.pnq.local/eng-build/iib:297697",
"locuz_version": "v4.9"
},
{
"added_bundle_images": [
"locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.7.7-14"
],
"index_image": "locuz.qe.pnq.local/eng-build/iib:297497",
"locuz_version": "v4.7"
},
{
"added_bundle_images": [
"locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.9.2-3"
],
"index_image": "locuz.qe.pnq.local/eng-build/iib:297495",
"locuz_version": "v4.9"
}
]
I have a ansible variable which contains one of the values from added_bundle_images. for example i have a variable called lastcsv which contains value "locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.7.7-14"
I am trying to write a json_query and also trying selectattr to get index_image if my key value matches "locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.7.7-14"
- name: set my variable
set_fact:
lastcsv: "locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.7.7-14"
- name: my debug2
debug: msg="{{ allcsv | selectattr('added_bundle_images','equalto', [lastcsv]) | list }}"
The output i see is :
TASK [common : my debug2] ***************************************************************************************************************************************************
ok: [foo.example.com] => {
"msg": []
}
Need help in constructing the query so that i get the index_image value when the added_bundle_image key has value "locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.7.7-14"
In allcsv, added_bundle_images is a list (denoted by []), so we can use contains match rather than equalto to see if the lastcsv element is in that list. Something like below:
- name: set my variable
set_fact:
lastcsv: "locuz.qe.pnq.local/eng-build/lzfirewallmod:v4.7.7-14"
- name: my debug2
debug:
msg: "{{ allcsv | selectattr('added_bundle_images', 'contains', lastcsv) | map(attribute='index_image') | list }}"
Produces:
TASK [my debug2] ********************************************************************************************************************************************
ok: [localhost] => {
"msg": [
"locuz.qe.pnq.local/eng-build/iib:297497"
]
}
I have to create a list based on a dictionary.
To get the element from the dictionary, I need to join "the server" + "the domain". The problem is that I have 3 different domains.
At the moment, I'm repeating the code to be able to use the 3 different domains.
- name: "Get server instances {{ ansible_fqdn }}"
set_fact:
app_ps_mon_list: "{{ app_ps_mon_list | default ([]) + [ app_instance ] }}"
vars:
app_instance: |
[Java,<event_type>]
java critical 1-
*ARGS {{item.value.INSTANCE_NAME.split("/")[1]}}
with_dict: '{{ server_instances[ansible_hostname + "<DOMAIN1>"] }}'
when: '{{server_instances[ansible_hostname + "<DOMAIN1>"] is defined and item.key != "SERVER_IMPACT"}}'
- name: "Get server instances {{ ansible_fqdn }}"
set_fact:
app_ps_mon_list: "{{ app_ps_mon_list | default ([]) + [ app_instance ] }}"
vars:
app_instance: |
[Java,<event_type>]
java critical 1-
*ARGS {{item.value.INSTANCE_NAME.split("/")[1]}}
with_dict: '{{ server_instances[ansible_hostname + "<DOMAIN2>"] }}'
when: '{{server_instances[ansible_hostname + "<DOMAIN2>"] is defined and item.key != "SERVER_IMPACT"}}'
- name: "Get server instances {{ ansible_fqdn }}"
set_fact:
app_ps_mon_list: "{{ app_ps_mon_list | default ([]) + [ app_instance ] }}"
vars:
app_instance: |
[Java,<event_type>]
java critical 1-
*ARGS {{item.value.INSTANCE_NAME.split("/")[1]}}
with_dict: '{{ server_instances[ansible_hostname + "<DOMAIN3>"] }}'
when: '{{server_instances[ansible_hostname + "<DOMAIN3>"] is defined and item.key != "SERVER_IMPACT"}}'
I've been trying to do the same with_subelements without success. Also I tried to use "ansible_fqdn", but the fqdn domain usually don't match the actual domain (I know its a mess).
Is there any workaround I could use to avoid repeating the code?
UPDATE
The idea of the playbook is to obtain the solutions from the server I'm using as host (ansible_hostname varaible).
Once it get the solutions, create a list using some information from the INSTANCE_NAME
This is a generic version of the dictionary:
{
"<server_name><domain>": {
"<solution_id>": {
"INSTANCE_NAME": "",
"SOLUTION_CATEGORY": "",
"SOLUTION_NAME": ""
},
"<solution_id>": {
"INSTANCE_NAME": "",
"SOLUTION_CATEGORY": "",
"SOLUTION_NAME": ""
},
SERVER_IMPACT: "",
...,
}
This is how the list should look like (its a multiline string variable, dumb but useful):
[
[Java]
java critical 1-
*ARGS <iINSTANCE_NAME info>
],
[
[Java]
java critical 1-
*ARGS <iINSTANCE_NAME info>
]
I have 2 different dictionaries that contains application information I need to join together.
landscape_dictionary:
{
"app_1": {
"Category": "application",
"SolutionID": "194833",
"Availability": null,
"Environment": "stage",
"Vendor/Manufacturer": null
},
"app_2": false
}
app_info_dictionary:
{
"app_1": {
"app_id": "6886817",
"owner": "owner1#nomail.com",
"prod": [
"server1"
],
"stage": []
},
"app_2": {
"app_id": "3415012",
"owner": "owner2#nomail.com",
"prod": [
"server2"
],
"stage": [
"server3"
]
}
}
This is the code I'm using to join both dictionaries
- set_fact:
uber_dict: "{{app_info_dictionary}}"
- set_fact:
uber_dict: "{{ uber_dict | default ({}) | combine(new_item, recursive=true) }}"
vars:
new_item: "{ '{{item.key}}' : { 'landscape': '{{landscape_dictionary[item.key]|default(false)}}' } }"
with_dict: "{{ uber_dict }}"
- debug:
msg: "{{item.key}}: {{item.value}}"
with_dict: "{{uber_dict}}"
If the value in the landscape_dictionary is false it will add it to the uber_dict without problems. But if the value contains information, it fails.
This is the error:
fatal: [127.0.0.1]: FAILED! => {"msg": "|combine expects dictionaries, got u\"{ 'app_1' : { 'landscape': '{u'Category': u'application', u'SolutionID': u'194820', u'Availability': None, u'Environment': 'stage', u'Vendor/Manufacturer': None}' } }\""}
What could be the problem?
Do I need to do an extra combine when I set the var in the set_fact?
Thanks
As #DustWolf notes in the comments,
For anyone from the Internet looking for the answer to: "How tp combine nested dictionaries in ansible", the answer is | combine(new_item, recursive=true)
This solves a closely related issue that has baffled myself and my team for months.
I will demonstrate:
Code:
---
- hosts: localhost
gather_facts: false
vars:
my_default_values:
key1: value1
key2:
a: 10
b: 20
my_custom_values:
key3: value3
key2:
a: 30
my_values: "{{ my_default_values | combine(my_custom_values, recursive=true) }}"
tasks:
- debug: var=my_default_values
- debug: var=my_values
Output:
ok: [localhost] =>
my_values:
key1: value1
key2:
a: 30
key3: value3
Note how key2 was completely replaced, thus losing key2.b
We changed this to:
my_values: "{{ my_default_values | combine(my_custom_values, recursive=true) }}"
Output:
my_values:
key1: value1
key2:
a: 30
b: 20
key3: value3
This syntax is not legal, or at the very least doesn't do what you think:
new_item: "{ '{{item.key}}' : { 'landscape': '{{landscape_dictionary[item.key]|default(false)}}' } }"
Foremost, ansible will only auto-coerce JSON strings into a dict, but you have used python syntax.
Secondarily, the way to dynamically construct a dict is not to use jinja2 to build up text but rather use the fact that jinja2 is almost a programming language:
new_item: "{{
{
item.key: {
'landscape': landscape_dictionary[item.key]|default(false)
}
}
}}"
Any time you find yourself with nested jinja2 interpolation blocks, that is a code smell that you are thinking about the problem too much as text (by nested, I mean {{ something {{nested}} else }})
I want to dynamically create a list of dictionaries that looks like this:
[ {'host': 'hostname1', 'id': 1}, {'host': 'hostname2', 'id': 2}, ]
And assign it to a variable in my playbook.
This variable is needed for a role I am using.
My attempt is the following:
- hosts:
- some-hosts
vars:
zk_hosts: []
tasks:
- name: create my var
set_fact:
zk_hosts: "{{ zk_hosts + [ {'host': item.1, 'id': item.0} ] }}"
with_indexed_items: "{{ groups.some-hosts }}"
However, when I run the playbook I have this warning:
[WARNING]: While constructing a mapping from stack.yml, line 16, column 3, found a duplicate dict key (vars). Using last defined value only.
And and error at this play:
fatal: [192.168.0.21]: FAILED! => {"failed": true, "msg": "ERROR! 'zk_hosts' is undefined"}
If I don't define zk_hosts before trying to set the fact, I get an error that the variable is undefined.
How can I solve?
EDIT
Easy fix, I just defined zk_hosts within the same task...
tasks:
- name: create my var
vars:
zk_hosts: []
set_fact:
zk_hosts: "{{ zk_hosts + [ {'host': item.1, 'id': item.0} ] }}"
with_indexed_items: "{{ groups.some-hosts }}"
Anyway, if there is a less cumbersome way of achieving the same, please advise!
You can use default filter:
set_fact:
zk_hosts: "{{ zk_hosts|default([]) + [ {'host': item.1, 'id': item.0} ] }}"
with_indexed_items: "{{ groups.some-hosts }}"