ansible/ansible-modules-extras

Ansible extra modules - these modules ship with ansible

ansible-modules-extras's Issues

Getent module's fail_key parameter does not work correctly with False or equivalents

Issue Type:

Bug Report

Ansible Version:

Both stable and devel:

ansible 1.8.2
configured module search path = None

ansible 1.9 (devel fe53d86) last updated 2015/01/06 08:46:10 (GMT +200)
lib/ansible/modules/core: (detached HEAD 0d551d8) last updated 2015/01/06 08:50:12 (GMT +200)
lib/ansible/modules/extras: (detached HEAD d4f5b6d) last updated 2015/01/06 08:50:12 (GMT +200)
v2/ansible/modules/core: (detached HEAD cb69744) last updated 2014/12/16 07:51:34 (GMT +200)
v2/ansible/modules/extras: (detached HEAD 8a4f07e) last updated 2014/12/16 07:51:45 (GMT +200)
configured module search path = None

Environment:

Running from Debian Testing.

Managing RHEL 6.6.

Summary:

Getent module fails looking up a nonexistent database entry when using fail_key=False parameter, while the expected behaviour is to succeed.

Steps To Reproduce:

Use getent to retrieve a nonexistent key in a database:

- name: getent test
  getent: database=passwd key=nonexistent fail_key=False

- debug: var=getent_passwd

Expected Results:

TASK: [getent test] ************************ 
ok: [<<REDACTED>>]

TASK: [debug var=getent_passwd] **********************************
ok: [<<REDACTED>>] => {
    "getent_passwd": {
        "nonexistent": null
    }
}

Actual Results:

TASK: [getent test] ************************ 
failed: [<<REDACTED>>] => {"failed": true}
msg: One or more supplied key could not be found in the database.

FATAL: all hosts have already failed -- aborting

Fix

The argument_spec entry for fail_key is missing "type='bool'":

diff --git a/system/getent.py b/system/getent.py
index 7da1be4..bb6d162 100644
--- a/system/getent.py
+++ b/system/getent.py
@@ -86,7 +86,7 @@ def main():
             database = dict(required=True),
             key      = dict(required=False, default=None),
             split    = dict(required=False, default=None),
-            fail_key = dict(required=False, default=True),
+            fail_key = dict(required=False, type='bool', default=True),
         ),
         supports_check_mode = True,
     )

Document whether resize2fs is used with lvol

Or any other filesystem resizing. It's not clear from the documentation whether I need to run this as a separate step or whether ansible does it automagically.

Group managment bug in FreeBSD (old #9513)

Issue Type:

"Group managment bug in FreeBSD"
Ansible Version:

ansible 1.7.2
Python 2.7.8
Environment:

--server
FreeBSD 9.3-RELEASE FreeBSD 9.3-RELEASE #0 r268512: Thu Jul 10 23:44:39 UTC 2014 [email protected]:/usr/obj/usr/src/sys/GENERIC amd64

--target_client
FreeBSD 10.0-RELEASE FreeBSD 10.0-RELEASE #0 r260789: Thu Jan 16 22:34:59 UTC 2014 [email protected]:/usr/obj/usr/src/sys/GENERIC amd64
Summary:

I try to change groups for user "test1" in FreeBSD host.
If I am adding a new groups to user (using module "user")- all is working fine , but when I try to "delete" some group for user "test1", I see status " : ok=2 changed=1 ", but realy nothing happens. And every time I see such status.
In Linux host all i working fine.
Please fix it.

--- First time (all is fine) ---

hosts: my_servers remote_user: root tasks:
name: Create FreeBSD users user: name={{ item.name }} password={{ item.password }} groups={{ item.groups }} comment={{ item.comment }} append=no with_items: - { name: 'test1' , groups: 'mysql,cyrus', password: '11111', comment: 'Vasja' } - { name: 'test2' , groups: 'mysql', password: '22222', comment: 'Kolja' } when: ansible_os_family == "FreeBSD"

--- Second time (bug) ----

hosts: my_servers remote_user: root tasks:
name: Create FreeBSD users user: name={{ item.name }} password={{ item.password }} groups={{ item.groups }} comment={{ item.comment }} append=no with_items: - { name: 'test1' , groups: 'mysql', password: '11111', comment: 'Vasja' } - { name: 'test2' , groups: 'mysql', password: '22222', comment: 'Kolja' } when: ansible_os_family == "FreeBSD"

root@kris:/home/serg/ansible # ansible-playbook -vvvv run.yml

PLAY [my_servers] *************************************************************

GATHERING FACTS ***************************************************************
ESTABLISH CONNECTION FOR USER: root
REMOTE_MODULE setup
EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/root/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'ConnectTimeout=10', '172.29.251.106', "/bin/sh -c 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1415297453.89-59397932578099 && echo $HOME/.ansible/tmp/ansible-tmp-1415297453.89-59397932578099'"]
PUT /tmp/tmpyPFA5r TO /root/.ansible/tmp/ansible-tmp-1415297453.89-59397932578099/setup
EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/root/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'ConnectTimeout=10', '172.29.251.106', u"/bin/sh -c 'LANG=en_US.UTF-8 LC_CTYPE=en_US.UTF-8 /usr/local/bin/python /root/.ansible/tmp/ansible-tmp-1415297453.89-59397932578099/setup; rm -rf /root/.ansible/tmp/ansible-tmp-1415297453.89-59397932578099/ >/dev/null 2>&1'"]
ok: [freebsd]

TASK: [Create FreeBSD users] **************************************************
ESTABLISH CONNECTION FOR USER: root
REMOTE_MODULE user name=test1 password=VALUE_HIDDEN groups=mysql comment=Vasja append=no
EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/root/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'ConnectTimeout=10', '172.29.251.106', "/bin/sh -c 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1415297454.25-201202690342400 && echo $HOME/.ansible/tmp/ansible-tmp-1415297454.25-201202690342400'"]
PUT /tmp/tmpDSK0ET TO /root/.ansible/tmp/ansible-tmp-1415297454.25-201202690342400/user
EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/root/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'ConnectTimeout=10', '172.29.251.106', u"/bin/sh -c 'LANG=en_US.UTF-8 LC_CTYPE=en_US.UTF-8 /usr/local/bin/python /root/.ansible/tmp/ansible-tmp-1415297454.25-201202690342400/user; rm -rf /root/.ansible/tmp/ansible-tmp-1415297454.25-201202690342400/ >/dev/null 2>&1'"]
changed: [freebsd] => (item={'comment': 'Vasja', 'password': '11111', 'name': 'test1', 'groups': 'mysql'}) => {"append": false, "changed": true, "comment": "Vasja", "group": 1002, "groups": "mysql", "home": "/home/test1", "item": {"comment": "Vasja", "groups": "mysql", "name": "test1", "password": "11111"}, "move_home": false, "name": "test1", "password": "NOT_LOGGING_PASSWORD", "shell": "/bin/sh", "state": "present", "uid": 1002}
ESTABLISH CONNECTION FOR USER: root
REMOTE_MODULE user name=test2 password=VALUE_HIDDEN groups=mysql comment=Kolja append=no
EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/root/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'ConnectTimeout=10', '172.29.251.106', "/bin/sh -c 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1415297454.45-101295486595285 && echo $HOME/.ansible/tmp/ansible-tmp-1415297454.45-101295486595285'"]
PUT /tmp/tmpSb6Sfc TO /root/.ansible/tmp/ansible-tmp-1415297454.45-101295486595285/user
EXEC ['ssh', '-C', '-tt', '-vvv', '-o', 'ControlMaster=auto', '-o', 'ControlPersist=60s', '-o', 'ControlPath=/root/.ansible/cp/ansible-ssh-%h-%p-%r', '-o', 'KbdInteractiveAuthentication=no', '-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no', '-o', 'ConnectTimeout=10', '172.29.251.106', u"/bin/sh -c 'LANG=en_US.UTF-8 LC_CTYPE=en_US.UTF-8 /usr/local/bin/python /root/.ansible/tmp/ansible-tmp-1415297454.45-101295486595285/user; rm -rf /root/.ansible/tmp/ansible-tmp-1415297454.45-101295486595285/ >/dev/null 2>&1'"]
+ok: [freebsd] => (item={'comment': 'Kolja', 'password': '22222', 'name': 'test2', 'groups': 'mysql'}) => {"append": false, "changed": false, "comment": "Kolja", "group": 1003, "groups": "mysql", "home": "/home/test2", "item": {"comment": "Kolja", "groups": "mysql", "name": "test2", "password": "22222"}, "move_home": false, "name": "test2", "password": "NOT_LOGGING_PASSWORD", "shell": "/bin/sh", "state": "present", "uid": 1003}

+PLAY RECAP ********************************************************************
freebsd : ok=2 changed=1 unreachable=0 failed=0
Steps To Reproduce:

Just run my playbook
ansible-playbook run.yml
Expected Results:

Please enter your expected results in this space. When running the steps supplied above in the previous section, what did you expect to happen? If showing example output, please indent your output by four spaces so it will render correctly in GitHub's viewer thingy.
Change user group for user "test1" from groups=1002(test1),88(mysql),60(cyrus) to groups=1002(test1),88(mysql)
Actual Results:

Still all old groups for user "test1"

uid=1002(test1) gid=1002(test1) groups=1002(test1),88(mysql),60(cyrus)

Bower and DNF modules not in PPA package

Issue Type:

Bug Report

Ansible Version:

ansible 1.8.2

Environment:

Ubuntu 14.10 (package installed using official PPA)

Summary:

The Bower and DNF modules are not included in the official PPA package.

Steps To Reproduce:

1. Install Ansible from PPA
2. dpkg -L ansible | grep 'bower\|dnf'

Expected Results:

/usr/share/pyshared/ansible/modules/extras/packaging/bower.py
/usr/share/pyshared/ansible/modules/extras/packaging/dnf.py

Actual Results:

Nothing - modules are not included in the distribution.

LVG module: Device xxx not found when using multiple devices

I'm getting the error: Device /dev/xvdb /dev/xvdc not found when using multiple devices. It appears to be a list being treated as a string separated by spaces. The problem appears to be in this line: https://github.com/ansible/ansible-modules-extras/blob/devel/system/lvg.py#L145

zypper module doesn't iterate over items

Issue Type:

Bug Report

Ansible Version:

ansible 1.8.2
configured module search path = /usr/share/ansible

Environment:

SLES 11 SP3

Summary:

I have some tasks that are supposed to install several packages via the zypper module. Since about Ansible 1.8 and specifically with version 1.8.2, the module appears to be broken, because it now tries to install a comma-separated list of modules and fails. It should install each package one-by-one.

Steps To Reproduce:

- name: install openmotif-libs packages
  zypper: name={{ item }} state=latest
  with_items:
  - openmotif-libs
  - openmotif-libs-32bit

Expected Results:

ok: [hostname] => (item=openmotif-libs) => {"changed": false, "item": "openmotif-libs", "msg": "remote module does not support check mode", "skipped": true} skipping: [hostname] => (item=openmotif-libs)
ok: [hostname] => (item=openmotif-libs-32bit) => {"changed": false, "item": "openmotif-libs-32bit", "msg": "remote module does not support check mode", "skipped": true} skipping: [hostname] => (item=openmotif-libs-32bit)

Actual Results:

failed: [hostname] => (item=openmotif-libs,openmotif-libs-32bit) => {"failed": true, "item": "openmotif-libs,openmotif-libs-32bit"} msg: No provider of 'openmotif-libs,openmotif-libs-32bit' found.

mention module development guide in readme (was: Need a simple way to test modules)

C:\ansible-modules-core\files>py stat.py
Traceback (most recent call last):
  File "stat.py", line 151, in <module>
    from ansible.module_utils.basic import *
ImportError: No module named ansible.module_utils.basic

Failed to validate the SSL certificate for Slack

Issue Type:

Bug Report

Ansible Version:

ansible 1.8

Environment:

OS X 10.10, Python 2.7.8 installed with homebrew.

Summary:

Certificate validation for Slack fails for some reason. It shouldn't, as the certificates are valid.

Steps To Reproduce:

Add a Slack notification to your playbook and play it:

    - name: Send notification to Slack
      local_action:
        module: slack
        domain: mediately.slack.com
        token: placeholder
        msg: "EC2 instance killed."

Expected Results:

A Slack notification.

Actual Results:

I get a certificate validation failure:

TASK: [Send notification to Slack] ******************************************** 
failed: [127.0.0.1 -> 127.0.0.1] => {"failed": true}
msg: Failed to validate the SSL certificate for mediately.slack.com:443. Use validate_certs=no or make sure your managed systems have a valid CA certificate installed. Paths checked for this platform: /etc/ssl/certs, /etc/ansible

FATAL: all hosts have already failed -- aborting

There are no certificates in /etc/ssl/certs, but the code adds a dummy certificate (https://github.com/ansible/ansible/blob/e54178f904d826904e2456bd8fa711d13480c5f7/lib/ansible/module_utils/urls.py#L220-L221), which should work out of the box?

Curl, wget both verify https://mediately.slack.com without a problem.

npm install returns "rc:1" but is instead successful

I'm trying to install a few npm modules (appium, cordova, ios-sim, ios-deploy) using the npm module.
All of them fail although when I check the npm packages are installed succesfully. As a result, running Ansible again results in everything green. Below is one output I got for appium.

failed: [192.168.1.28] => (item={'version': '1.2.0', 'name': 'appium'}) => {"cmd": "/usr/local/bin/npm install --global [email protected]", "failed": true, "item": {"name": "appium", "version": "1.2.0"}, "rc": 1}
stderr:

stdout:
> [email protected] install /usr/local/lib/node_modules/appium/node_modules/ws
> (node-gyp rebuild 2> builderror.log) || (exit 0)

Could this be because Ansible is sensitive about the error output from npm or could this be npm problem? I tried extracting the run_command method from AnsibleModule and run it separately and it exit successfully.

Add support for "cpanm --mirror-only"

Issue Type:

Feature Idea

Ansible Version:

1.7.2

Environment:

Ubuntu, but this goes for any environment.

Summary:

It would be great if it was possible to specify mirror_only=yes in the cpanm module. This enable installs from minicpan in a local environment where extra packages has been injected.

Example:

cpanm --mirror-only --mirror http://cpan.myinternal.net My::Custom::Module

Without "--mirror-only", cpanm will try to lookup My::Custom::Module in CPAN Meta DB.

Steps To Reproduce:

Expected Results:

Actual Results:

installing the latest npm with the npm module doesn't work

Issue Type:

Bug Report

Component Name:

npm

Ansible Version:

ansible 1.7.1

Environment:

OSX 10.10

Summary:

Attempting to install the latest npm with the npm module doesn't seem to work.

Steps To Reproduce:

Put the following in a playbook where node is installed

- name: make sure the latest npm is installed
  npm: name=npm global=true

...or try running the following:
tkellen/ansible-vagrant-example@903ee91

Expected Results:

Latest version of npm is installed.

Actual Results:

NPM version is unchanged.

lvg handling multiple drives as one

Also see: #56

It seems this has come back to us.

vgcreate/extend command returns:
Unable to add physical volume '/dev/xvdb /dev/xvdc' to volume group 'vg.osddata'

while the playbook sent:

REMOTE_MODULE lvg vg=vg.osddata state=present pvs=/dev/xvdb,/dev/xvdc

I tried with this version:
https://raw.githubusercontent.com/ansible/ansible-modules-extras/devel/system/lvg.py

Thanks,
Mark

monitoring/monit.py won't recognize the name of a process which contains uppercase letters

Issue Type:

Bug Report

Ansible Version:

1.7.2 (tested)
branch: devel (i looked into the code, see Possible Solution)

Environment:

Centos 6.6

Summary:

monitoring/monit.py won't recognize a process name if the name contains uppercase letters.
sorry, but that's all.

Steps To Reproduce:

- name: myMonitProcess
  monit: name=myMonitProcess state=unmonitored

will result in

msg: myMonitProcess process not presently configured with monit

if "monit summary" is displaying:

Process 'myMonitProcess'  Running
System 'localhost' Running

Expected Results:

I expect it to work with uppercase process names as it is working with lowercase only process names, and mixedcase process names self speaking.

Actual Results:

Unless i use completely lowercase process names it won't work.

Possible Solution:

Sorry, i don't have a test environment, but i managed to identify a possible solution:

line 78 and line 79

parts = line.lower().split()
if len(parts) > 2 and parts[0] == 'process' and parts[1] == "'%s'" % name:

'line' should NOT be converted to lower and 'process' should then be 'Process' (uppercase P)

The alternatives module is erroneously reporting "changed" status.

Issue Type:

Bug Report

Ansible Version:

ansible 1.8 (devel ba46930) last updated 2014/08/06 15:37:06 (GMT -600)

Environment:

CentOS 6

Summary:

The alternatives module is reporting "changed" status when re-running a task that has already been successfully applied.

Steps To Reproduce:

Given a play:

---
- hosts: all
  sudo: yes
  vars:
    jdk_dir: /usr/java/jdk1.7.0_55

  tasks:
  - name: Configure Java alternatives
    alternatives: name={{ item.name }} link={{ item.link }} path={{ item.path }}
    with_items:
    - { name: jar, link: /usr/bin/jar, path: "{{ jdk_dir }}/bin/jar" }
    - { name: jps, link: /usr/bin/jps, path: "{{ jdk_dir }}/bin/jps" }
    - { name: java, link: /usr/bin/java, path: "{{ jdk_dir }}/bin/java" }
    - { name: jmap, link: /usr/bin/jmap, path: "{{ jdk_dir }}/bin/jmap" }
    - { name: javac, link: /usr/bin/javac, path: "{{ jdk_dir }}/bin/javac" }
    - { name: javaws, link: /usr/bin/javaws, path: "{{ jdk_dir }}/bin/javaws" }
    - { name: jstack, link: /usr/bin/jstack, path: "{{ jdk_dir }}/bin/jstack" }
    - { name: jvisualvm, link: /usr/bin/jvisualvm, path: "{{ jdk_dir }}/bin/jvisualvm" }

Expected Results:

The first time the task is run the status is "changed" as expected. We expect subsequent runs with the identical path specified to indicate an "ok" status.

Actual Results:

When running the task subsequently the status is actually reporting "changed".

Note: This ticket was previously reported as ansible #8613.

ufw module does not allow specifying both interface and source IP

It looks as though the ufw module does not allow specifying both an interface and a from_ip. The following code block was producing what appears to have been a ufw-related error:

- name: Allow MySQL/Galera Traffic Over Cluster Network
  ufw: rule=allow proto=tcp
        from_ip="{{ mysql_galera_network }}"
        interface="{{ mysql_galera_interface }}"
        to_port={{ item }}
  with_items:
    - "4567"
    - "3306"
  when: mysql_use_galera == true

The resulting error is:

...
failed: [stage-mysqlcluster4.teamsnap.com] => (item=3306) => {"failed": true, "item": "3306"} msg: ERROR: Invalid token 'on'
...

Removing the interface line, however, alleviates the issue. The documentation on the module page (http://docs.ansible.com/ufw_module.html) indicates this should be allowed, although the examples do not necessarily match the options table.

zypper module doesn't respect disable_gpg_check=yes option for state=latest

Issue Type:

Bug Report

Ansible Version:

ansible 1.5.4
Based on code inspection, this is reproducible on the development branch as well.

Environment:

Ansible Host: Ubuntu 14.04.1 LTS
Target System: SUSE 11.3 (zypper 1.6.308)

Summary:

Using the following Ansible script:

zypper: name=package-name state=latest disable_gpg_check=yes

A previously-installed package will not be updated to the latest version, if the repository serving the update fails a certificate check. It is expected that the disable_gpg_check=yes flag would bypass this, but it does not appear to work properly.

Steps To Reproduce:

Install an out-of-date package on the target system and check its version.
Create an Ansible task:

- name: install the zypper repository
  zypper_repository: name=myrepo repo=http://path_to_repo/ state=present

- name: update the component
  zypper: name=package-name state=latest disable_gpg_check=yes

Create the zypper repository, populate the repository with a newer version of the package, and ensure that the repository is unsigned.
Run the playbook.
Check the version of the package that's installed on the target system.

Expected Results:

It's expected that the component's package will be updated to the newest version, and that will be reflected in Ansible's output:

TASK: [component | install the zypper repository] ***************************** 
ok: [192.168.0.14]

TASK: [component | update the component] ************************************** 
changed: [192.168.0.14]

Actual Results:

The component's package remains at the old version, and there is no error or other indication of this in Ansible's output:

TASK: [component | install the zypper repository] ***************************** 
ok: [192.168.0.14]

TASK: [component | update the component] ************************************** 
ok: [192.168.0.14]

Notes:

In the zypper module it seems that the package_latest() function doesn't respect disable_gpg_check=yes like package_present() does.

This indicates that Ansible will generate this command:

/usr/bin/zypper --non-interactive update --auto-agree-with-licenses package-name

when it should generate this command:

/usr/bin/zypper --non-interactive --no-gpg-check update --auto-agree-with-licenses package-name

If I run the commands manually, the first fails as expected with:

File 'repomd.xml' from repository 'myrepo' is unsigned, continue? [yes/no] (no): no
Retrieving repository 'myrepo' metadata [error]
Repository 'myrepo' is invalid.

mongodb_user works incorrectly with old pymongo

Issue Type: Bug report

Ansible Version: 1.7.1

Environment:

OSX 10.7 -> Ubuntu precise

Summary:

mongodb_user works incorrectly with old pymongo (~2.1).
Tested with 2.7.2, work correctly.

Steps To Reproduce:

• Fresh mongo instance with auth enabled, no users
• ansible host -m mongodb_user -a "database=admin user=test roles=root password=test state=present"

Expected Results:

Creates new user

Actual Results:

Changed = false

This code block supposed to fail on non-existant user:

        try:
           client[db_name].authenticate(user, password)
           if state == 'present':
              module.exit_json(changed=False, user=user)
        except OperationFailure:
           if state == 'absent':
              module.exit_json(changed=False, user=user)

But it is exiting correctly and continue.

Old issue

ansible/ansible#8899

lvol module doesn't take options for lvextend/lvreduce and lvreduce uses --force by default

Hello,

We heavily use the lvol/lvg modules in our setups.
We've noticed that to shrink a volume we need to pass the --force option which works.
However, after a few ansible-playbook runs we've noticed that somehow one volume had been resized (we can't prove if ansible is responsible) hence its associated mountpoint could not be accessed anymore since we should have done a resizefs prior to it: we ended up re creating the filesystem and lost some data.

We would like to prevent the loss of data by specifying to the lvol module (extend or reduce but not lvcreate) the --resizefs option like this lvreduce_options=--resizefs (and lvextend_options).

Specifying --resizefs to lvreduce acts as a safe guard to the data integrity of the volume since depending on the number of files/inodes on the device, the call to resizefs will fail if the size is smaller than the threshold below which data will get corrupted.

we've noticed that lvreduce is always called with --force which is not good guys! Please don't use --force by default, let the user specify it via lvreduce_options.

If the lvm design is done right, shrinking is just a matter of: lvreduce --resizefs -L XG
So no --force and in addition the FS is handled prior to the lvm shrinking, all good.

Of course this is less of a problem in case of lvextend but again passing --resizefs will make sure the FS will get expanded too after the lvm expansion which a lot of users tend to forget.

thanks for the good work

Ansible 1.8.2, module: lvg. Creation of volume group with symlink in pvs list is not idempotent.

Issue Type:

Bug Report

Ansible Version:

1.8.2

Environment:

Ansible on Gentoo amd64.

Summary:

Creation of volume group with symlink in pvs list is not idempotent. When symlink is in pvs list should check vg members with real devices not link name.

Steps To Reproduce:

ls -l /dev/md/raid1
lrwxrwxrwx 1 root root 8 Jan 12 18:26 /dev/md/raid1 -> ../md127

Repeat the task at least twice:

• name: create raid1 vg

  lvg: vg=raid1 pvs=/dev/md/raid1

  tags: storage

Expected Results:

Create at the first run, ok on the second and next runs.

Actual Results:

TASK: [create raid1 vg] *******************************************************
failed: [s1.ifab.ru] => {"err": " Can't initialize physical volume "/dev/md/raid1" of volume group "raid1" without -ff\n", "failed": true, "rc": 5}
msg: Creating physical volume '/dev/md/raid1' failed

npm module traceback

this playbook:

• hosts: all
  tasks:
  • name: install global npm packages
    npm: name={{ item }} global=yes state=latest
    with_items:
    • gulp

gives this error:

TASK: [install global npm packages] *******************************************
failed: [localhost] => (item=gulp) => {"failed": true, "item": "gulp", "parsed": false}
Traceback (most recent call last):
File "/Users/Morriz/.ansible/tmp/ansible-tmp-1416254319.77-187721358027197/npm", line 1828, in
main()
File "/Users/Morriz/.ansible/tmp/ansible-tmp-1416254319.77-187721358027197/npm", line 248, in main
installed, missing = npm.list()
File "/Users/Morriz/.ansible/tmp/ansible-tmp-1416254319.77-187721358027197/npm", line 167, in list
data = json.loads(self._exec(cmd, True, False))
File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/json/init.py", line 338, in loads
return _default_decoder.decode(s)
File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/json/decoder.py", line 365, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/json/decoder.py", line 383, in raw_decode
raise ValueError("No JSON object could be decoded")
ValueError: No JSON object could be decoded

slack 404

I'm getting the same issue as was reported here: ansible/ansible#8669

Error output I'm getting is:

failed: [test-domain.com -> 127.0.0.1] => {"failed": true}
msg:  failed to send payload={"username": "toby", "text": "stage update has successfully completed", "link_names": 1, "channel": "#deploys", "icon_url": "http://www.ansible.com/favicon.ico"} to https://pixelfusion.slack.com/services/hooks/incoming-webhook?token=[obscured]: HTTP Error 404: Not Found

subversion should not check revision info after exporting repository.

If you export repository, you get clean directory structure without .svn folder. And if you try to make svn info in that folder - you get an error. So ansible should not check revision info after exporting repository

SQL Syntax error with mysql_replication

Issue Type:

Bug Report

Ansible Version:

ansible 1.8.2
configured module search path = /usr/share/ansible

Environment:

OS X 10.10.1 > Debian 7.7

Summary:

Hello, I try to setup my slaves using ansible, but I have an issue while setting the master state into the slave. I get a mysql exception regarding an SQL Syntax Error.

Steps To Reproduce:

- name: get master replication status
  mysql_replication: mode=getmaster login_user={{ master.user }} login_password={{ master.password }}
  delegate_to: "{{master.host}}"
  register: replication
  when: slave|failed

- name: define master status in slave
  mysql_replication: mode=changemaster master_host={{ master.host }} master_log_file={{ replication.File }} master_log_pos={{ replication.Position }} master_user={{ master.user }} master_password={{ master.password }}
  when: slave|failed

Expected Results:

No error

Actual Results:

TASK: [db-slave | define master status in slave] *****************************
failed: [a-future-slave] => {"failed": true, "parsed": false}
Change master
Traceback (most recent call last):
  File "/root/.ansible/tmp/ansible-tmp-1418728948.93-89965021745143/mysql_replication", line 1949, in <module>
    main()
  File "/root/.ansible/tmp/ansible-tmp-1418728948.93-89965021745143/mysql_replication", line 367, in main
    changemaster(cursor, chm, chm_params)
  File "/root/.ansible/tmp/ansible-tmp-1418728948.93-89965021745143/mysql_replication", line 165, in changemaster
    cursor.execute(query, chm_params)
  File "/usr/lib/python2.7/dist-packages/MySQLdb/cursors.py", line 174, in execute
    self.errorhandler(self, exc, value)
  File "/usr/lib/python2.7/dist-packages/MySQLdb/connections.py", line 36, in defaulterrorhandler
    raise errorclass, errorvalue
_mysql_exceptions.ProgrammingError: (1064, "You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near ''130779818'' at line 1")

From the a-future-slave logs, here is the output I get :

Dec 16 12:22:32 a-future-slave ansible-mysql_replication: Invoked with master_ssl_cert=None master_log_pos=130779818 master_ssl=False master_log_file=mysql-bin.000450 login_user=None login_host=localhost master_port=None master_host=my-master master_ssl_ca=None master_ssl_key=None login_unix_socket=None master_connect_retry=None master_user=my_user master_ssl_capath=None mode=changemaster relay_log_pos=None login_password=NOT_LOGGING_PASSWORD master_ssl_cipher=None master_password=my_password relay_log_file=None

Firewalld should accept an 'enabled' flag

Issue Type:

Feature Idea

Ansible Version:

ansible 1.8.2

Environment:

Centos 7

Summary:

The firewalld module follows the command structure very closely, ie: you need to call it twice to make a permanent change, and a change in the current firewall rules.

- name: Add keepalived service to internal zone (permanent)
  firewalld: zone=internal service=keepalived state=enabled permanent=true

- name: Add keepalived service to internal zone (now)
  firewalld: zone=internal service=keepalived state=enabled permanent=false

The module would be easier to use if one could use an 'enabled' flag:

- name: Add keepalived service to internal zone (permanent, now)
  firewalld: zone=internal service=keepalived state=enabled permanent=true enabled=true

- name: Add keepalived service to internal zone (only on next restart, not now)
  firewalld: zone=internal service=keepalived state=enabled permanent=false enabled=false

- name: Add keepalived service to internal zone (not permanent, now)
  firewalld: zone=internal service=keepalived state=enabled permanent=false enabled=true

The final possible combination (not permanent or now) should trigger an error.

I very rarely make a change to the firewall that I don't want to have happen immediately AND on reboot.

Thank you so much for the excellent module!

Composer module won't work as expected

On ansible 1.7.2 whenever I try to use composer: command=install working_dir=/path/to/project the task fails:

TASK: [Running composer install] **********************************************
failed: [default] => {"failed": true}
msg: unsupported parameter for module: command

FATAL: all hosts have already failed -- aborting

firewalld: will remove previos rules

firewalld will only have the latest rule applied to the settings.

If you run this Task:

• name: Firewall settings
  firewalld: zone=public port=8080/tcp permanent=true state=enabled
  firewalld: zone=public port=8983/tcp permanent=true state=enabled
  firewalld: zone=public port=443/tcp permanent=true state=enabled
  firewalld: zone=public port=80/tcp permanent=true state=enabled
• shell: firewall-cmd --reload

You get this result with misses ports.

[root@localhost ~]# firewall-cmd --list-all
public (default)
interfaces:
sources:
services: dhcpv6-client ssh
ports: 443/tcp 80/tcp
masquerade: no
forward-ports:
icmp-blocks:
rich rules:

This is wrong, but as a workaround you can call firewalld from Shell and it works well.

• name: Firewall settings
  shell: firewall-cmd --permanent --zone=public --add-port=8080/tcp
  shell: firewall-cmd --permanent --zone=public --add-port=8983/tcp
  shell: firewall-cmd --permanent --zone=public --add-port=80/tcp
  shell: firewall-cmd --permanent --zone=public --add-port=443/tcp
  shell: firewall-cmd --reload

Slack webhook notification does not work with ansible 1.8

Issue Type:

Bug Report

Ansible Version:

Broken : ansible 1.8.1 ansible-1.8.1-1.el6.noarch
Working : ansible 1.7.2 ansible-1.7.2-2.el6.noarch

Environment:

CentOS release 6.6 (Final)

Summary:

Slack webhook notification does not work anymore with the update of ansible 1.8.1.
Tasks says there is no problem but nothing is getting shown in slack.

Steps To Reproduce:

- name: Send notification message via Slack all options
  local_action:
    module: slack
    domain: xxx.slack.com
    token: xxx
    msg: "Branch {{ build }}, is now getting build on jenkins with these options: {{params}} optimize:{{optimize}} bootstrap:{{bootstrap}} and will stay up for {{timetolive}} days. The process can be followed here : xxx"
    channel: "#ci"
    link_names: 1

Expected Results:

A post to the slack channel:
"Branch test, is now getting build on jenkins with these options: test optimize:test bootstrap:test and will stay up for test days. The process can be followed here : xxx"

Actual Results:

Nothing is shown.

1.7.2 -vvv result (working)

TASK: [Send notification message via Slack all options] ***********************
<127.0.0.1> REMOTE_MODULE slack channel='#ci' token= domain=xxx.slack.com msg='Branch test, is now getting build on jenkins with these options: test optimize:test bootstrap:test and will stay up for test days. The process can be followed here : xxxx'
<127.0.0.1> EXEC ['/bin/sh', '-c', 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1418310211.18-188493459373960 && chmod a+rx $HOME/.ansible/tmp/ansible-tmp-1418310211.18-188493459373960 && echo $HOME/.ansible/tmp/ansible-tmp-1418310211.18-188493459373960']
<127.0.0.1> PUT /tmp/tmplhxI3I TO /home/ansible/.ansible/tmp/ansible-tmp-1418310211.18-188493459373960/slack
<127.0.0.1> EXEC /bin/sh -c 'sudo -k && sudo -H -S -p "[sudo via ansible, key=uhkhypxfjylljezednmskwqgaokaogyq] password: " -u root /bin/sh -c '"'"'echo SUDO-SUCCESS-uhkhypxfjylljezednmskwqgaokaogyq; LANG=C LC_CTYPE=C /usr/bin/python /home/ansible/.ansible/tmp/ansible-tmp-1418310211.18-188493459373960/slack; rm -rf /home/ansible/.ansible/tmp/ansible-tmp-1418310211.18-188493459373960/ >/dev/null 2>&1'"'"''
ok: [localhost -> 127.0.0.1] => {"changed": false, "msg": "OK"}

1.8.1 -vvv result (broken)

TASK: [Send notification message via Slack all options] ***********************
<127.0.0.1> REMOTE_MODULE slack channel='#ci' token= domain=xxx.slack.com msg='Branch test, is now getting build on jenkins with these options: test optimize:test bootstrap:test and will stay up for test days. The process can be followed here : xxx'
<127.0.0.1> EXEC ['/bin/sh', '-c', 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1418310038.98-146454958098479 && chmod a+rx $HOME/.ansible/tmp/ansible-tmp-1418310038.98-146454958098479 && echo $HOME/.ansible/tmp/ansible-tmp-1418310038.98-146454958098479']
<127.0.0.1> PUT /tmp/tmpRiLiit TO /home/ansible/.ansible/tmp/ansible-tmp-1418310038.98-146454958098479/slack
<127.0.0.1> EXEC /bin/sh -c 'sudo -k && sudo -H -S -p "[sudo via ansible, key=ciahqceudvthumdpkvppcfzugjjtmefp] password: " -u root /bin/sh -c '"'"'echo SUDO-SUCCESS-ciahqceudvthumdpkvppcfzugjjtmefp; LANG=C LC_CTYPE=C /usr/bin/python /home/ansible/.ansible/tmp/ansible-tmp-1418310038.98-146454958098479/slack; rm -rf /home/ansible/.ansible/tmp/ansible-tmp-1418310038.98-146454958098479/ >/dev/null 2>&1'"'"''
ok: [localhost -> 127.0.0.1] => {"changed": false, "msg": "OK"}

Add support of "interface direct" on firewalld

ISSUE TYPE

Feature Idea

COMPONENT NAME

firewalld module

ANSIBLE VERSION

N/A

SUMMARY

hi guys,

that would be nice if the firewalld module could support the "interface direct" configuration, allowing us to run this kind of rules:
firewall-cmd --permanent --direct --add-chain ipv4 filter FORWARD -i krb0 -o krb0 -j ACCEPT

locale_gen does not generate locales on Arch Linux

Issue Type:

Bug report

Ansible Version:

1.6.10

Environment:

Arch Linux (both master and slave)

Summary:

locale_gen doesn't generate locales on Arch Linux.

The locale_gen module looks for lines starting with a hash and a space, and removes the hash to enable locale generation for those lines. But Arch's locale.gen doesn't have spaces (only the hash) so no lines get uncommented. Since no locales are enabled, locale-gen does nothing.

Steps To Reproduce:

Create a fresh Arch Linux system and install openssh/python2 on it. Add it to your Ansible hosts.

$ ansible -m locale_gen -a "name=en_US.UTF-8 state=present" --extra-vars "ansible_python_interpreter=/usr/bin/python2" <your Arch Linux target here>
$ ssh <your arch linux target here> locale -a

Expected Results:

The Ansible ad-hoc command should succeed and indicate "changed"
The ssh command locale -a should list all the locales on the system. en_US.UTF-8 should be on the list.

Actual Results:

The Ansible command does succeed, but C and POSIX are the only locales listed aftewards. en_US.UTF-8 does not appear.

Note: clone of ansible/ansible#8387

lvg fails when vg_options is not supplied

Issue Type: Bug

Ansible Version: ansible 1.6.6

Environment: Ubuntu 14.04

Summary: lvg module fails when vg_options is not supplied

Steps To Reproduce:

ansible -i inventory host -m lvg -a "vg='cinder' pvs='/dev/sda5'"

Expected Results: Volume group created

Actual Results:

Volume group is not created as the module attempts to split on a parameter that hasn't been supplied, note that the docs specify this parameter as optional.

hugh@host:~/$ ansible -i inventory node18 -m lvg -a "vg='cinder' pvs='/dev/sda5'"
node18 | FAILED >> {
    "failed": true,
    "msg": "Traceback (most recent call last):\n  File \"<stdin>\", line 1458, in <module>\n  File \"<stdin>\", line 119, in main\nAttributeError: 'NoneType' object has no attribute 'split'\n",
    "parsed": false
}

Zypper should be able to install "pattern"

The ansible zypper module is currently not able to install a "pattern".

Issue Type:

“Feature Idea”

Ansible Version:

1.8.1

Environment:

openSUSE 13.1

Summary:

Use case on the shell, that is not covered by the zypper module:
zypper install -t pattern kde

This could be implemented by another argument like
type='package', 'pattern' which is then passed to zypper.
There are other types like 'srcpackage', 'product', 'patch' which zypper knows of.

The get_current_version and get_package_state functions needs to be adjusted to correctly treat patterns.

Steps To Reproduce:

zypper: type=pattern name=kde state=present

Expected Results:

installs the pattern kde

Actual Results:

not possible

homebrew_cask should accept an appdir parameter

Issue Type:

Feature Idea.

Ansible Version:

ansible --version
ansible 1.8.2
configured module search path = None

Environment:

I'm running Ansible on OSX 10.10.1 and I'm managing the same computer (local connection).

Summary:

By default, cask symlinks the apps to ~/Applications instead of /Applications.

The brew cask command accepts a "--appdir" parameter but with homebrew_cask it is impossible to set it. So I would like to request an improvement to the module to add this parameter.

At the moment I've patched locally the source of the module by hardcoding it to /Applications like this:

    cmd = [opt
           for opt in (self.brew_path, 'cask', 'install', '--appdir=/Applications', self.current_cask)
           if opt]

(lines 398-400 in homebrew_cask.py). It seems to work so far. But this is not the right solution, obviously.

Steps To Reproduce:

For example, the homebrew_cask task could look like this:

- name: Install Homebrew cask apps
  homebrew_cask: "name=spotify state=present appdir=/Applications"

Alternatives module: "'NoneType' object has no attribute 'startswith'"

Issue Type:

Bug Report

Ansible Version:

1.7.2

Environment:

• Centos 5.5 (python 2.4.3 & python 2.6.8)
• Centos 6.5 (python 2.6.6)

Summary:

Alternatives module generates invalid python code on Centos 5 & 6
The error seems to be more related to the version of python than the OS as I get the same Centos 6 error on Centos 5 when specifying ansible_python_interpreter in the /etc/ansible/hosts

Steps To Reproduce:

    - yum: name=java-1.6.0-openjdk-devel state=installed
    - alternatives: name=java path=/usr/lib/jvm/jre-1.6.0-openjdk.x86_64/bin/java

Expected Results:

Install java 1.6 and then configures the alternatives to point /usr/bin/java to the 1.6 version

Actual Results:

fatal: [Centos6] => failed to parse: Traceback (most recent call last):
  File "/tmp/.ansible/tmp/ansible-tmp-1412275620.84-98969527997889/alternatives", line 1486, in <module>
    main()
  File "/tmp/.ansible/tmp/ansible-tmp-1412275620.84-98969527997889/alternatives", line 122, in main
    check_rc=True
  File "/tmp/.ansible/tmp/ansible-tmp-1412275620.84-98969527997889/alternatives", line 1339, in run_command
    args = [ os.path.expandvars(os.path.expanduser(x)) for x in args ]
  File "/usr/lib64/python2.6/posixpath.py", line 251, in expanduser
    if not path.startswith('~'):
AttributeError: 'NoneType' object has no attribute 'startswith'

failed: [Centos5] => {"failed": true, "parsed": false}
invalid output was: Traceback (most recent call last):
  File "/tmp/.ansible/tmp/ansible-tmp-1412275620.83-139068779696675/alternatives", line 1486, in ?
    main()
  File "/tmp/.ansible/tmp/ansible-tmp-1412275620.83-139068779696675/alternatives", line 132, in main
    except subprocess.CalledProcessError, cpe:
AttributeError: 'module' object has no attribute 'CalledProcessError'


FATAL: all hosts have already failed -- aborting

When using python26 on Centos5:

fatal: [Centos5] => failed to parse: Traceback (most recent call last):
  File "/tmp/.ansible/tmp/ansible-tmp-1412275580.11-200946814475830/alternatives", line 1486, in <module>
    main()
  File "/tmp/.ansible/tmp/ansible-tmp-1412275580.11-200946814475830/alternatives", line 122, in main
    check_rc=True
  File "/tmp/.ansible/tmp/ansible-tmp-1412275580.11-200946814475830/alternatives", line 1339, in run_command
    args = [ os.path.expandvars(os.path.expanduser(x)) for x in args ]
  File "/usr/lib64/python2.6/posixpath.py", line 251, in expanduser
    if not path.startswith('~'):
AttributeError: 'NoneType' object has no attribute 'startswith'

I also get the same No module named yum as reported in this question

Hipchat V2

The Hipchat module only supports API V1 at this moment. Anyone working on API V2 or interested in it?
Ideally, something like this:

$ hipchat: token=AAAAAA room=notify msg="Ansible task finished" hipchat_api="v2"

By declare hipchat_api we can force API v2 and give backward compatible for API v1. No change needs to make to existed playbook. So it won't break anything.

at module: work on FreeBSD

Issue Type:

Bug Report

Ansible Version:

ansible 1.8.2
configured module search path = /usr/local/share/ansible

Environment:

FreeBSD

Summary:

at.py constructs a command line for the "at" utility with positional arguments last.
Glibc allows this, but FreeBSD libc (and others) do not. This is a security feature.

So, this needs to be written as "at -f /tmp/cmd now + 1 minute" to be portable to non-glibc systems.

--- at.py.orig  2014-12-15 11:08:47.153125860 -0500
+++ at.py   2014-12-15 11:09:11.137125129 -0500
@@ -78,7 +78,7 @@


 def add_job(module, result, at_cmd, count, units, command, script_file):
-    at_command = "%s now + %s %s -f %s" % (at_cmd, count, units, script_file)
+    at_command = "%s -f %s now + %s %s" % (at_cmd, script_file, count, units)
     rc, out, err = module.run_command(at_command, check_rc=True)
     if command:
         os.unlink(script_file)

Steps To Reproduce:

On FreeBSD, do: ansible -m at -a 'command="touch /tmp/t" count=2 units="minutes"' 127.0.0.1

Expected Results:

There is a /tmp/t file touched two minutes from "now"

Actual Results:

127.0.0.1 | FAILED >> {
    "cmd": "/usr/bin/at now + 2 minutes -f /tmp/at47k09b", 
    "failed": true, 
    "msg": "at: garbled time", 
    "rc": 1, 
    "stderr": "at: garbled time\n", 
    "stdout": ""
}

Virt module: get_xml missing flags option

Issue Type:

Feature Request

Ansible Version:

ansible 1.7.1

Environment:

N/A

Summary:

The get_xml method in the virt module has the flags parameter hardcoded to 0, which prevents passing in a custom flags setting to for example also retrieve the VNC password used.

Steps To Reproduce:

Running the following does not show the password of the domain in the XML:
ansible -sK -m virt -a "command=get_xml name=[VM name]"

Expected Results:

The domain XML including the VNC password when flasg is set to 1.
Retrieving the password of a virtual domain is currently not possible, would be nice to be able to call:
ansible -sK -m virt -a "command=get_xml name=[VM name] xml_secure=yes"

Actual Results:

Adding extra parameters causes the module to fail due to an unrecognised parameter

Use -z when invoking pkg_add to allow installation of alternatives

Issue Type:

Bug Report/Feature Request
(WAS ansible/ansible#8990)

Ansible Version:

1.6.1, but same in master

Environment:

OpenBSD

Summary:

openbsd_pkg 's module doesnt allow to install a package when it has alternatives - ie for example php-fpm, which could be 5.3 or 5.4 - or python, where 2.6, 2.7 and 3.3 are available

Steps To Reproduce:

playbook snippet:

- name: install php-fpm
  openbsd_pkg: name=php-fpm-5.4 state=installed

Expected Results:

php-fpm 5.4 should have installed

Actual Results:

failed: [twilight.rhaalovely.net] => {"failed": true, "item": ""}
msg: Error from http://ftp.fr.openbsd.org/pub/OpenBSD/5.5/packages/amd64/php-fpm-5.4.tgz
ftp: Error retrieving file: 404 Not Found

pkg_add has this feature that if you use -z, it will try fuzzy-matching the package name/version, if you provided it enough details

Various usescases for -z ..

$ sudo pkg_add -I php-fpm-5.4*  
Error from http://ftp.fr.openbsd.org/pub/OpenBSD/5.5/packages/amd64/php-fpm-5.4*.tgz
ftp: Error retrieving file: 404 Not Found
$ sudo pkg_add -Iz php-fpm      
Ambiguous: php-fpm could be php-fpm-5.3.28p3 php-fpm-5.4.24
$ sudo pkg_add -Iz php-fpm-5.4*
php-fpm-5.4.24: ok

If openbsd_pkg would add -z in pkg_add invocation in https://github.com/ansible/ansible/blob/devel/library/packaging/openbsd_pkg#L127, then i would be able to install the correct version of the package within the alternatives, specifying php-fpm-5.4* in the name arg.

[Slack] Be able to configure the full slack hook URL

Hi,

I was trying to add Slack notification if one of my Ansible script and the Slack configuration I have got for an Incoming Webhook does not contains any Token information.

Instead, Slack provide me with a full URL (Probably unique for my webhook) that has nothing to do with what the current Slack Module is trying to build.

It look like that:
https://hooks.slack.com/services/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

If the Slack team is really changing the way hook integrations are working, it would probably be more future proof to allow Ansible users to set the full URL instead (or on top) of the Token.

Aurélien

zypper_repository module should have a fingerprint option

When adding a repository using zypper_repository, its public key may not automatically be stored as trusted. In case of the "zypper" command, it would ask interactively whether a certain GPG key fingerprint would be accepted by the user or not. However, when scripting this using the ansible zypper_repository module, this is not possible.

Some users use the "disable_gpg_check" option, but this disables the GPG check completely, thus opening a security vulnerability.

Thus, the user of the ansible zypper_repository module should be able to explicitly specify an acceptable GPG key (or multiple acceptable GPG keys) by verbatimly quoting the fingerprint of that key. This way, it is prevented that untrusted software is installed using ansible.

Example

Instead of

zypper_repository: repo=http://download.opensuse.org/repositories/Application:/Geo/openSUSE_13.1/ name=/Application:/Geo/openSUSE_13.1/ state=present disable_gpg_check=yes

use

zypper_repository: repo=http://download.opensuse.org/repositories/Application:/Geo/openSUSE_13.1/ name=/Application:/Geo/openSUSE_13.1/ state=present acceptable_gpg_key_fingerprint=195E211106BC205D2A9C2222CC7F07489591C39B

The pkgutil provider reports state: present when a package is not in the catalog

When I try to install a non-existing package with something like

# ansible unstable9s -m pkgutil -a "name=CSWlibpsl-dev state=present"

The result is success and the state "present" is reported:

unstable9s | success >> {
    "changed": true, 
    "name": "CSWlibpsl-dev", 
    "state": "present", 
    "stderr": "Package CSWlibpsl-dev not in catalog. Exiting.\n", 
    "stdout": "Solving needed dependencies ...\n"
}

The package was not installed on the machine and is not installed afterwards:

unstable9s% pkginfo -x CSWlibpsl-dev
ERROR: information for "CSWlibpsl-dev" was not found

bigip_facts: get_fw_rule NotImplemented Error

Getting the following error after a recent F5 upgrade to 11.6 from 11.5

failed: [localhost -> 127.0.0.1] => {"failed": true}
msg: received exception: Server raised fault: 'Exception caught in LocalLB::urn:iControl:LocalLB/VirtualServer::get_fw_rule()
Common::NotImplemented

Role is defined as the following:

  tasks:
  - name: Get F5 Virtual Servers
    local_action: >
      bigip_facts
      server=esilo-f5-corp
      user={{ f5_user }}
      password={{ f5_passwd }}
      include=virtual_server
    register: bigip_vars

Ansible 1.7.2

rabbitmq_user fails when no password provided

Even though the password param is specified as optional, the rabbitmq_user module fails when it is not provided.

Stacktrace:
  File "<stdin>", line 1595, in <module>
  File "<stdin>", line 240, in main
  File "<stdin>", line 165, in add
  File "<stdin>", line 131, in _exec
  File "<stdin>", line 1448, in run_command
  File "/usr/lib/python2.7/posixpath.py", line 261, in expanduser
    if not path.startswith('~'):
AttributeError: 'NoneType' object has no attribute 'startswith'

The bug is right here. I might send a PR when I get the time to fork and test.

RFE: Docker Facts module

ISSUE TYPE:

• Feature Idea

COMPONENT NAME:

• docker_facts

ANSIBLE VERSION:

1.8

ENVIRONMENT:

Ubuntu 13.04

SUMMARY:

Docker facts module. This module returns facts pertaining to a running fleet of containers as well as images.

It would be useful to have a docker_facts module to return information about running containers as well I found an issue with the docker module that I ended up discussing with Paul Durivage. The essential problem is that there is a list of dictionaries per container that only contains the last running container when using "with_sequence" or "with_items" versus "count", hence making it impossible to utilize information for the entirety of a running fleet of docker containers outside of a task using the docker module.

An idea to deal with this short-coming would be to have a "docker_facts" module for obtaining information from the fleet of Docker containers.

STEPS TO REPRODUCE:

N/A

EXPECTED RESULTS:

Dictionary containing two primary entries for containers and images

ACTUAL RESULTS:

N/A

mkfs.xfs uses -f instead of -F flag

Issue Type:

Bugfix Pull Request

Ansible Version:

1.7.2

Environment:

Ubuntu 14.04

Summary:

mkfs.xfs uses the -f instead of -F flag to force creating a filesystem on devices that have an existing filesystem #18

Steps To Reproduce:

filesystem: fstype=xfs dev=/dev/sdb1 force=yes

Expected Results:

device formatted as xfs

Actual Results:

xfs uses the -f flag to force creating a filesystem on a devices that have an existing filesystem; however, -F flag is being passed, which mkfs.xfs doesn't accept.

Add the "name" parameter in logentries follow module

Issue Type:

“Feature Idea”

Ansible Version:

1.7.2

Environment:

N/A

Summary:

In the current module logentries , it will only accept the path parameter to following logs.

In logentries CLI agent, you can now add in a name parameter to name your

$ le follow '/var/log/nginx*.log' --name="nginx_logs"

Steps To Reproduce:

when we do

- logentries: path=/var/log/nginx*.log state=present

On logentries, it defaults the name of this tracking to "_.log" instead for the path /var/log/nginx_.log. This makes it slightly confusing and not nice looking (harder to organize).

Expected Results:

Propose to add a name parameter instead so that we can do

- logentries: path=/var/log/nginx*.log state=present name=nginx_logs

Actual Results:

N/A

zabbix_maintenance : Incorrect method "user.authenticate"

Issue Type:

Bug Report

Ansible Version:

ansible --version
ansible 1.8
configured module search path = /usr/share/ansible

Environment:

CentOS release 6.5 (Final)
Zabbix Server 2.4.2

Summary:

zabbix_maintenance Invalid params., Incorrect method "user.authenticate"

Steps To Reproduce:

ansible-playbook zabbix-test.yml -vvvv

zabbix-test.yml :

  - name: Create 30 minutes maintenance window in Zabbix for the Web group
    zabbix_maintenance: name="Web release"
                      host_groups=Web
                      state=present
                      minutes=30
                      server_url=http://zabbix.example.net
                      login_user=zabbixuser
                      login_password=zabbixpassword

Expected Results:

Zabbix maintenance to be created successfully in the frontend.

Actual Results:

TASK: [Create 30 minutes maintenance window in Zabbix for the Web group] ****
<localhost> ESTABLISH CONNECTION FOR USER: zabbixuser on PORT 22 TO localhost
<localhost> REMOTE_MODULE zabbix_maintenance name="Web release" host_groups=CERES_Prod state=present minutes=30 server_url=http://zabbix.example.net login_user=zabbixuser login_password=VALUE_HIDDEN
<localhost> EXEC /bin/sh -c 'mkdir -p $HOME/.ansible/tmp/ansible-tmp-1417009357.56-165948609075556 && chmod a+rx $HOME/.ansible/tmp/ansible-tmp-1417009357.56-165948609075556 && echo $HOME/.ansible/tmp/ansible-tmp-1417009357.56-165948609075556'
<localhost> PUT /tmp/tmpBh9O6y TO /home/zabbixuser/.ansible/tmp/ansible-tmp-1417009357.56-165948609075556/zabbix_maintenance
<localhost> EXEC /bin/sh -c 'sudo -k && sudo -H -S -p "[sudo via ansible, key=kjuyjmwmetzdqhinxvpkhxmgpqgkaroo] password: " -u root /bin/sh -c '"'"'echo SUDO-SUCCESS-kjuyjmwmetzdqhinxvpkhxmgpqgkaroo; LANG=C LC_CTYPE=C /usr/bin/python /home/zabbixuser/.ansible/tmp/ansible-tmp-1417009357.56-165948609075556/zabbix_maintenance; rm -rf /home/zabbixuser/.ansible/tmp/ansible-tmp-1417009357.56-165948609075556/ >/dev/null 2>&1'"'"''
failed: [localhost] => {"failed": true}
msg: Failed to connect to Zabbix server: (u'Error -32602: Invalid params., Incorrect method "user.authenticate". while sending {"params": {"password": "zabbixpassword", "user": "zabbixuser"}, "jsonrpc": "2.0", "method": "user.authenticate", "auth": "", "id": 0}', -32602)

FATAL: all hosts have already failed -- aborting

Notes:

According to the latest Zabbix version documentation (https://www.zabbix.com/documentation/2.4/manual/api/reference/user) there doesn't seem to be any user.authenticate method supported but it should be user.login .

zfs create fails when parent fs doesn't exist

msg: cannot create 'tank/deleteme/us/nagios1.lup1/blah': parent does not exist

zfs list
tank 1.94T 95.8T 663G none
tank/deleteme 472K 95.8T 236K none
tank/deleteme/us 236K 95.8T 236K none

I did try to lookup the Zfs class and there is no use of '-p' (i am using FreeBSD)
zfs create [-pu] [-o property=value]... filesystem
I tried changing the below

def create(self):
    if self.module.check_mode:
        self.changed = True
        return
    properties=self.properties
    volsize = properties.pop('volsize', None)
    volblocksize = properties.pop('volblocksize', None)
    if "@" in self.name:
        action = 'snapshot'
    else:
       action = 'create' 

tried to change the above to action = 'create -p'
but is not working.

__init__.py missed in windows directory

as result, the *.py files don't installed and those plugins can't be detected.

Add install instructions to README

Issue Type: Documentation Report

Ansible Version: 1.8.2

Environment: N/A

Summary: Add install instructions to README. Neither the documentation nor the README of this repo mention how to install the extras-modules.

Steps To Reproduce: Go to the README, modules docs, dev docs and try to find installation instructions for extras-modules.

Expected Results: The README should have an "Installation" chapter describing how to install the extras-modules (e.g. "pip install ansible-modules-extras" if there were such a pip package).

Actual Results: (see "Steps To Reproduce" above) -- nowhere installation of modules is mentioned.

ec2_vpc module always returns "changed = true"

Issue Type:

Bug Report

Ansible Version:

ansible-playbook 1.8 (devel 3bd1e4f) last updated 2014/08/28 21:51:11 (GMT -400)

Environment:

boto 2.3.2.1

Summary:

Running the same VPC creation tasks always results in status changed. This seems to be triggered by line 492 of ec2_vpc, but that is an indicator that something is happening that shouldn't be.

Steps To Reproduce:

- hosts: localhost
  gather_facts: no
  tasks:
  - name: create vpc
    ec2_vpc:
      state: present
      cidr_block: 172.22.0.0/16
      resource_tags: { "Name": "TEST" }
      subnets:
        - cidr: 172.22.1.0/24
          resource_tags: { "Name":"TEST-1" }
          az: us-east-1b
        - cidr: 172.22.2.0/24
          resource_tags: { "Name":"TEST-2" }
          az: us-east-1c 
      route_tables:
        - subnets:
            - 172.22.1.0/24
            - 172.22.2.0/24
          routes:
            - dest: 0.0.0.0/0
              gw: igw 
      internet_gateway: True
      region: "{{ region }}"
      wait: yes
    register: vpc

Expected Results:

Creation would happen on the first time, but subsequent times changed should be = False.

Actual Results:

first run

$: ansible-playbook -i inventory/hosts test.yml -v

PLAY [localhost] **************************************************************

TASK: [create vpc] ************************************************************
changed: [localhost] => {"changed": true, "subnets": [{"az": "us-east-1c", "cidr": "172.22.2.0/24", "id": "subnet-45f92c32", "resource_tags": {"Name": "TEST-2"}}, {"az": "us-east-1b", "cidr": "172.22.1.0/24", "id": "subnet-0b2f3823", "resource_tags": {"Name": "TEST-1"}}], "vpc": {"cidr_block": "172.22.0.0/16", "dhcp_options_id": "dopt-d2cadfb0", "id": "vpc-0c5fea69", "region": "us-east-1", "state": "available"}, "vpc_id": "vpc-0c5fea69"}

PLAY RECAP ********************************************************************
localhost                  : ok=1    changed=1    unreachable=0    failed=0

second run

$: ansible-playbook -i inventory/hosts test.yml -v

PLAY [localhost] **************************************************************

TASK: [create vpc] ************************************************************
changed: [localhost] => {"changed": true, "subnets": [{"az": "us-east-1c", "cidr": "172.22.2.0/24", "id": "subnet-45f92c32", "resource_tags": {"Name": "TEST-2"}}, {"az": "us-east-1b", "cidr": "172.22.1.0/24", "id": "subnet-0b2f3823", "resource_tags": {"Name": "TEST-1"}}], "vpc": {"cidr_block": "172.22.0.0/16", "dhcp_options_id": "dopt-d2cadfb0", "id": "vpc-0c5fea69", "region": "us-east-1", "state": "available"}, "vpc_id": "vpc-0c5fea69"}

PLAY RECAP ********************************************************************
localhost                  : ok=1    changed=1    unreachable=0    failed=0