* Adds another module utility file which generalizes the
access of urls via the urllib* libraries.
* Adds a new spec generator for common arguments.
* Makes the user-agent string configurable.
Fixes#6211
0. Uncomment the test.
1. Test fails.
2. Make vars unique per file in test inventory files.
3. Modify token addition to not ast.literal_eval(v) a variable containing a hash.
4. Modify vars to have an escape in test inventory file.
5. Catch exceptions explicitly. Any unknown exceptions should be a bug.
6. Test passes.
Allow security tokens and profiles to be used as arguments
to the 'common' ec2 modules
Mostly refactoring to provide two new methods,
`get_aws_connection_info`, which results in a dict that can be
passed through to the boto `connect_to_region` calls, and
`connect_to_aws` that can pass that dict through to the
`connect_to_region` method of the appropriate module.
Tidied up some variable names
Works around boto/boto#2100
profiles don't work with boto < 2.24, but this detects for that
and fails with an appropriate message. It is designed to work
if profile is not passed but boto < 2.24 is installed.
Modifications to allow empty aws auth variables to be passed
(this is useful if wanting to have the keys as an optional
parameter in ec2 calls - if set, use this value, if not set,
use boto config or env variables)
Reworked validate_certs improvements to work with refactoring
Added documentation for profile and security_token to affected modules
In order to simplify the workflow with the GCE modules, it's now
possible to add the parameters and project name as arguments to the
various GCE modules.
The inventory plugin also returns the IP of the host in
`ansible_ssh_host` so that you don't have to specify IPs into the
inventory file.
Some update to the documentation are also added.
Closes#5583.
It came up that fixing this unit test may relate to another ticket that is open. This work allows us to uncomment this unit test by fixing how we pars variables allowing a quoted variable to contain a '#'.
Work also went into cleaning up some of the test data to clarify what was working.
Lastly work went into cleaning up formatting so that the code is easily read.
Add support for checking host against global known host files.
The effect of this is that before this fix if files are spread across the known_hosts file but not in the ~/known_hosts file the hosts will execute sequentially. This PR augments the functionality so that all of the knowns hosts will execute in parallel.
##### Issue Type:
Bugfix Pull Request
##### Ansible Version:
ansible 1.4.3
##### Environment:
N/A
##### Summary:
We are using a wrapper python script to run ansible-playbook. We use subprocess to execute and print the stdout as and when its written. Problem is when we use pause it doesn't display the prompt string as raw_input does not flush stdout before reading from stdin.
It looks like a dirty fix to add "\n" to the prompt string but i don't see any other way to over come this. If anyone else have a better fix please do propose/suggest.
##### Steps To Reproduce:
```yaml
#File: test_play.yml
- name: Test
hosts: $nodes
gather_facts: false
tasks:
- name: Waiting for User
local_action: pause prompt="Do you want to continue (yes/no)? "
```
```python
#!/usr/bin/env python
#File: test.py
import shlex, subprocess
def run_process(process):
process = process.encode("utf-8")
command = shlex.split(process)
p = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
for line in iter(p.stdout.readline, b''):
print line,
cmd = "/usr/bin/python -u /usr/bin/ansible-playbook -i hosts.txt test_play.yml -e 'nodes=local'"
run_process(cmd)
```
```
shell $ python test.py
```
##### Expected Results:
```
PLAY [Test] *******************************************************************
TASK: [Waiting for User] ******************************************************
[localhost]
Do you want to continue (yes/no)? :
```
##### Actual Results:
```
PLAY [Test] *******************************************************************
TASK: [Waiting for User] ******************************************************
[localhost]
```
Addresses GH-5116.
It comes up that when parsing json that if you are missing the last double quote on the last variable and the next line is just '}' we will get an out of range error. In this instance we will also then make sure that the line is long enough to have two colons.
Create a lookup plugin named dict that can be used to loop over hashes.
It converts a dict into a list of key-value pairs, with attributes named
"key" and "value." Also adds a brief explanation and simple example to
the docs.
Signed-off-by: Kent R. Spillner <kspillner@acm.org>
Bugfixes:
* the remote_src param was not being converted to a boolean correctly,
resulting in it never being used by the module as the default behavior
was remote_src=True (issue #5581)
* the remote_src param was not listed in the generic file params, leading
to a failure when the above bug regarding remote_src was fixed
* the delimiter should always end with a newline to ensure that the file
fragments do not run together on one line
Fixes#5581
This occurred when a hash would be passed in via extra args and the
hash variable behavior was set to 'merge', which resulted in the
variable from extra args replacing the playbook variable.
When content is processed and found to be valid JSON it is decoded into a dict. To write it out to a file we need to encode it back into a string.
Addresses GH-5914.
When disabled, the boto connection will be instantiated without validating
the SSL certificate from the target endpoint. This allows the modules to connect
to Eucalyptus instances running with self-signed certs without errors.
Fixes#3978
There is a bit going on with the changes here. Most of the changes are cleanup of files so that they line up with the standard files.
PR #5136 was merged into the current devel and brought up to working order. A few bug fixes had to be done to get the code to test correctly. Thanks out to @pib!
Issue #5431 was not able to be confirmed as it behaved as expected with a sudo user.
Tests were added via a playbook with archive files to verify functionality.
All tests fire clean including custom playbooks across multiple linux and solaris systems.
It turns out that some of the assumptions in #5885 were slightly off. The previous fix relied on a call to the module to creat a tmp_path. This is insufficent as there are few cases that we need to have the tmp directory before we make the module call. If we don't have a tmp_path before we do a recursive call or when we find a file that does not match the remote md5 hash we need to create a tmp directory. Also we are not more percise when we will need to clean up the remote tmp_path.
