In Ansible to use facts about a host you must have either:
already gathered facts from that host in that play (just gathering facts
would be enough, no tasks needed)
- hosts: webservers:dbservers
tasks: []
- hosts: webservers # that need to know facts about dbservers
tasks:
- ...
OR
enabled fact caching and previously talked to that host within the cache
timeout
http://docs.ansible.com/playbooks_variables.html#fact-caching
On Thu, Aug 28, 2014 at 3:25 PM, Abe Voelker <[email protected]> wrote:
> So I wrote a template task that populates this jinja template:
>
> {{ hostvars['198.58.107.247'] }}
>
> And this is the output I got:
>
> {'ansible_ssh_user': 'root', 'inventory_hostname_short': '198',
> 'group_names': ['webservers'], 'inventory_hostname': '198.58.107.247'}
>
> So hostvars['198.58.107.247'] doesn't seem to be filled with all the
> facts for that host. Then I realized that the root playbook I'm running (
> provision.yml) looks like this:
>
> ---
> - include: redisservers.yml
> - include: webservers.yml
>
> And the redisservers.yml playbook it includes looks like this:
>
> ---
> - hosts: redisservers
> roles:
> - redis
>
> And the *full* output I'm seeing from running the playbook looks like
> this (my emphasis):
>
> PLAY [redisservers]
> ***********************************************************
>
> *GATHERING FACTS
> *************************************************************** *
> *ok: [72.14.178.81]*
>
> TASK: [redis | Allow port 6379 (Redis) access to Web servers]
> *****************
> fatal: [72.14.178.81] => One or more undefined variables: 'dict object'
> has no attribute 'ansible_eth0'
>
> FATAL: all hosts have already failed -- aborting
>
> So from the output it looks like it has only gathered facts for
> 72.14.178.81 by the time the redisservers playbook runs. Therefore I
> fixed my issue by changing the order of provision.yml, which runs the
> webservers playbook first and populates the facts for 198.58.107.247 before
> the redisservers playbook runs:
>
> ---
> - include: webservers.yml
> - include: redisservers.yml
>
> Then I figured out that if I add a hosts: all line to provision.yml, it
> will gather facts for all hosts before running the included playbooks:
>
> ---
> - hosts: all
> - include: webservers.yml
> - include: redisservers.yml
>
> PLAY [all]
> ********************************************************************
>
> *GATHERING FACTS
> *************************************************************** *
> *ok: [72.14.178.81]*
> *ok: [198.58.107.247]*
>
> PLAY [webservers]
> *************************************************************
> ...
>
> So that way I don't have worry about the order of playbook inclusion. But
> this still seems kind of hacky... wouldn't it be better if there were a way
> to access hostvars in a manner that will just populate it for a host if it
> hasn't been populated yet? Sort of a lazy evaluation?
>
> On Thursday, August 28, 2014 12:19:24 PM UTC-5, Abe Voelker wrote:
>>
>> I have an inventory file like this:
>>
>> [redisservers]
>> 72.14.178.81
>>
>> [webservers]
>> 198.58.107.247
>>
>> I'm trying to run this playbook against the redisservers group to allow
>> access to port 6379 for the webservers:
>>
>> ---
>> - name: Allow port 6379 (Redis) access to Web servers
>> ufw: rule=allow interface=eth0 direction=in port=6379 proto=tcp src="{{
>> hostvars[item]['ansible_eth0']['ipv4']['address'] }}"
>> with_items: groups['webservers']
>>
>> But I'm getting this error:
>>
>> TASK: [redis | Allow port 6379 (Redis) access to Web servers]
>> *****************
>> fatal: [72.14.178.81] => One or more undefined variables: 'dict object'
>> has no attribute 'ansible_eth0'
>>
>> FATAL: all hosts have already failed -- aborting
>>
>> I'm not sure what's going on as all servers definitely have an eth0
>> interface. When I run ansible all -m setup -i inventory, the output
>> seems to confirm that there is an ansible_eth0 fact available for all
>> servers: https://gist.github.com/abevoelker/7a3c18f10793a744e6da
>>
>> I've tried simplifying the playbook to refer to the exact server in
>> question to remove the looping:
>>
>> ---
>> - name: Allow port 6379 (Redis) access to Web servers
>> ufw: rule=allow interface=eth0 direction=in port=6379 proto=tcp src="{{
>> hostvars[groups['webservers'][0]]['ansible_eth0']['ipv4']['address'] }}"
>>
>> and also
>>
>> ---
>> - name: Allow port 6379 (Redis) access to Web servers
>> ufw: rule=allow interface=eth0 direction=in port=6379 proto=tcp src="{{
>> hostvars['198.58.107.247']['ansible_eth0']['ipv4']['address'] }}"
>>
>> But I'm still getting the same error.
>>
>> Any hints as to what I'm doing wrong or how I can continue to debug this?
>> I'm using Ansible 1.7.1.
>>
> --
> You received this message because you are subscribed to the Google Groups
> "Ansible Project" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/ansible-project/792ddf1e-fdbf-430e-bb60-7647a554bbcb%40googlegroups.com
> <https://groups.google.com/d/msgid/ansible-project/792ddf1e-fdbf-430e-bb60-7647a554bbcb%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
> For more options, visit https://groups.google.com/d/optout.
>
--
You received this message because you are subscribed to the Google Groups
"Ansible Project" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/ansible-project/CA%2BnsWgxSJ-4X1ODff_DpjAUeiUs2vDf-OW09LPEK3R8svBGMkw%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.
