Is it possible to pass a list as an extra-vars variable on the command line?
I have tried various syntax but it doesn't work.
E.g.
ansible-playbook my_playbook.yml -i local --extra-vars
roles_to_deploy=role1,role2
ansible-playbook my_playbook.yml -i local --extra-vars
roles_to_deploy=[role1,ro
Hi Jeff,
I can only agree with Michael here - playbooks describe ordering of
operations for a specific operational purpose and as far as I know role
dependencies will not get you there in the scenario you describe.
The way I usually do this is to think of splitting my roles into two. You
generall
Thanks, I was trying to replace the default file with 0440 rights, thats
why i got this error. I copied it with different filename in same path,
works like a charm!
On Monday, March 16, 2015 at 4:17:17 PM UTC+2, tkuratomi wrote:
>
> This sounds like some sort of bug or misconfiguration. Ansibl
What happens if you reboot the box? I haven't tried this myself but I've
seen windows delay applying updates until shutdown (and I think following a
reboot too in some cases).
Might need a bit of work judging by the comments but there is a win_reboot
module which you could try.
https://github.
Hi, Jon.
I appreciate the response, and certainly will look into your suggestions.
I'm using version 1.8.4.
As I've continued to hack away, I've got this to work, kind of:
ansible somehost -m win_updates -a category:security
The output looks like this:
somehostb | success >> {
"changed":
Hi Dimitri,
A few others have reported StackOverflow and OutOfMemory exceptions on
unpatched Server 2008 machines -
see https://github.com/ansible/ansible/pull/8345#issuecomment-52074837 for
a hotfix that worked for me.
By the way you can put any custom modules you want to use in your
/etc/a
apt repos allow this but as far as I can tell it does not work for
PPAs, we can have diff package per arch or target distro .. but not
sure if that will even be possible as a work around for this.
--
Brian Coca
--
You received this message because you are subscribed to the Google Groups
"An
After unsuccessfully looking around for instructions on the internet I
took a look at what other PPAs were doing. I'm not a Ubuntu/Debian
user but I wasn't able to find any PPAs that have multiple versions
(there are packages for multiple ubuntu releases but not multiple
versions of the package fo
Hi Frederick/Dylan
I have the same problem. The repo query is being made to every single build
of a package, thus making the yum install several mins.
I did set the keepcache to 1 in /etc/yum.conf. but that did not fix the
issue. Could you share what those settings exactly are?
Thanks
Bala
I was actually planning to soften the restriction and allow -K to work with -b
--
Brian Coca
--
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 ansible-proj
We still haven't been able to find a PPA setting to disable the auto
delete (or to restore the previous version? There must be a toplevel
menu for configuring PPAs that we just aren't seeing...). As a
temporary solution you can grab the old .deb from here:
https://launchpad.net/~ansible/+archive/u
I usually use ansible with the "-sK" flags (use sudo, and ask for sudo
password). Now I've noticed the new "-b" flag. So I wanted to use it, but
I'm told I can't use "-b" and "-K" together. So I have to type:
-b --ask-become-pass
Is there a plan to introduce a short form for --ask-become-pass?
Okay, great. Thanks for the clarification, Brian.
On Thursday, March 26, 2015 at 10:12:57 AM UTC-5, Brian Coca wrote:
>
> No, the 1.8.4 issue was not expected, we are looking into it, it seems
> that previous versions were just removed, 1.8.4 SHOULD remain
> available.
>
>
>
> --
> Brian Coca
Let me try to better illustrate my question with a more concrete example.
Imagine an app Foo that has two tiers: Apache and MySQL. Each tier is a
role with start and stop tasks. These tasks are tagged appropriately. Foo
lists Apache and MySQL as its contained roles.
Starting the app is as simpl
Ansible 1.8.4 is still available, we've just always had our pip settings
configured to "hide" older releases. You can still install the older
version by specifying it to pip:
pip install ansible==1.8.4
James Cammarata
Director, Ansible Core Engineering
github: jimi-c
On Thu, Mar 26, 2015 at 10
On Thu, Mar 26, 2015 at 7:29 AM, Rob Zienert wrote:
> With the regressions in 1.9.0 related to the s3 module[0], I think the
> Ansible package repos really need to keep more than the HEAD stable version
> in its versions list. When 1.9.0 was released yesterday, all of my company's
> playbooks brok
No, the 1.8.4 issue was not expected, we are looking into it, it seems
that previous versions were just removed, 1.8.4 SHOULD remain
available.
--
Brian Coca
--
You received this message because you are subscribed to the Google Groups
"Ansible Project" group.
To unsubscribe from this group a
Also should have included this in the initial postI would prefer to not
have to use the synchronize module as that requires rsync on both machines.
On Thursday, March 26, 2015 at 10:59:16 AM UTC-4, Mark Raider wrote:
>
> I am having difficulty figuring out how to have ansible manage the cont
So, in a normal usecase, 1.8.4 would've remained available? My concern
isn't 1.9.0 vs 1.9.0.1, but that the last minor version is kept available.
On Thursday, March 26, 2015 at 10:00:29 AM UTC-5, Brian Coca wrote:
>
> The issue with 1.9.0 and 1.9.0.1 was caused by not being able to
> upload a fi
The issue with 1.9.0 and 1.9.0.1 was caused by not being able to
upload a file with the same version to pypy, initially an incomplete
tarball was uploaded, once we realized the error and deleted it we
were forced to change the version number to be able to release as pypy
does not allow overwriting
I am having difficulty figuring out how to have ansible manage the content
of a directory, including subdirectories.
For example..
if I have the following
example_top_level_dir/
-> subdir1
-> subdir2
And I remove the subdir2 from Ansible...I want that reflected on the remote
mach
I'm getting an unbalance quote error on this short playbook. I'm not
seeing the error
---
- hosts: devcluster
remote_user: ansible
sudo: yes
tasks:
-name: add jmx host entry to devjboss cluster
action: lineinfile dest="/etc/hosts" backup=yes state="present"
insertafter="mydomain
With the regressions in 1.9.0 related to the s3 module[0], I think the
Ansible package repos really need to keep more than the HEAD stable version
in its versions list. When 1.9.0 was released yesterday, all of my
company's playbooks broke since we make extensive use of S3 - so right now
we can
All,
I've got Ansible working on a few Windows 2008 servers. The Ansible
command server runs CentOS 6. I'm trying to get the module "win_updates"
working, but am not having much success. I'd like to be able to run this
both via the command line, and via a playbook. The extant documentation
Thanks, Dan!
I've managed to do what I wanted with you help.
I ended up with:
setting hash_behaviour = merge in ansible.cfg
group_vars/all
users:
user1:
comment: "user1"
state: "present"
groups:
- "wheel"
- "dvp"
authorized:
- "ssh-rsa 111"
- "ssh-rsa 222"
I guess there is a possibility that we create our own
loader/scanner/parser. But in the end, I don't really care, I don't find
it hard to remember to wrap ninja statements in quotes.
On another note, this is more suited for ansible-devel and not
ansible-project.
On Thu, Mar 26, 2015 at 8:35 AM,
I've seen a few threads about looping over a list on a role, and the
currently preferred method seems to be to pass the list as a variable and
do the looping in each task.
My problem is that I not only want to use each variable as is, but also
construct new variables based on logic over multipl
Is any work beeing done on enabling running a task in a remote namespace
(setns)?
Reason I am asking is I have a network oob namespace sshd is running in,
and for
some of the ansible tasks I would like beeing able to specify for example:
- namespace: netns=default
..which should cause all comma
I agree that forking is not a very good option for such a little addition,
though the YAML Python library hasn't been updated for 6 years, so there's
no risk missing a big upstream update anyway.
On the other hand, your answer is not really helpful here. What would be
the optimal solution to av
for tower related questions please email supp...@ansible.com
On Thu, Mar 26, 2015 at 8:53 AM, Nils De Moor wrote:
> Hi,
>
> For quite a while we are happy users of Ansible Tower.
> Now, we are refactoring our playbooks and roles and want to make everything
> more modular, by leveraging galaxy and
Hi,
For quite a while we are happy users of Ansible Tower.
Now, we are refactoring our playbooks and roles and want to make everything
more modular, by leveraging galaxy and having separate repositories per
role and a main repository with a requirements file.
As far as we can test, Tower does
I would saying forking the YAML module is an absolutely terrible idea.
On Thursday, March 26, 2015, Warren Seine wrote:
> Hi,
>
> Ansible handles pretty well the incompatibility between YAML and Jinja
> templates with a nice, informative, explicit error message. I suppose this
> is so because wr
Thanks for the hint Serge - you put me on the right track to the solution.
As you rightly said there was more needed in the yml file
Here is the working one.
---
- hosts: all
sudo: yes
tasks:
- name: restart server
command: /sbin/reboot
async: 0
poll: 0
ignore_err
Hi,
Ansible handles pretty well the incompatibility between YAML and Jinja
templates with a nice, informative, explicit error message. I suppose this
is so because writing foo: {{ bar }} is an extremely common pattern and
most users face the error. The workaround is simple but still, annoying.
34 matches
Mail list logo