with the default umask tar will create a world-readable archive of the
container, which may contain sensitive data
Signed-off-by: Evgeni Golov <evgeni@golov.de>
* do not use a predictable filename for the LXC attach script
* don't use predictable filenames for LXC attach script logging
* don't set a predictable archive_path
this should prevent symlink attacks which could result in
* data corruption
* data leakage
* privilege escalation
otherwise deploying user-containers fail as these require information
from ~/.config/lxc/default.conf that the LXC tools will load if no
--config was supplied
Signed-off-by: Evgeni Golov <evgeni@golov.de>
This change adds a note to the win_scheduled_task module
docs that indicates Windows Server 2012 or later is required.
This is because the module relies on the Get-ScheduledTask
cmdlet, which is a part of the Server 2012 OS. Previous
versions, like Server 2008, simply can't work with this
module.
The range_search() API was added to the shade library in version
1.5.0 so let's check for that and let the user know they need to
upgrade if they try to use it.
Addition of an os_ironic_inspect module to leverage the OpenStack
Baremetal inspector add-on to ironic or ironic driver out-of-band
hardware introspection, if supported and configured.
The manual check to see if get_bin_path() returned anything is
redundant, because we pass True to the required parameter of
get_bin_path(). This automatically causes the task to fail if the pacman
binary isn't available. Therefore, the code within the if statement
being removed is never called.
-e or --execute [1] allows to execute a specific piece of Puppet code
such a class.
For example, in puppet you would run:
puppet apply -e 'include ::mymodule'
Will be in ansible:
puppet: execute='include ::mymodule'
[1] http://docs.puppetlabs.com/puppet/latest/reference/man/apply.html#OPTIONS
win_unzip fails to extract files when either src or dest contains
complex paths such as "..\..\" or "C:\\Program Files" (double slashes).
Fix this by fetching absolute path of both before invoking CopyHere
method.
Set int for the various port (and so avoid to convert them later)
Set no_log=True for the login_password
Verify that db is a int, so avoid a conversion
Do a sorted comparison of the list of security groups supplied via `module.params.get('security_groups')` and the list of security groups fetched via `get_sec_group_list(eni.groups)`. This fixes an incorrect "The specified address is already in use" error if the order of security groups in those lists differ.
I changed the logic here to always use 'netsh ... show rule' keywords as keys for $fwsettings map. While the translation (e.g. Enabled -> enable) is performed when invoking 'netsh ... add rule' command.
I tested rule creation and rule creation when the rule was already existing on Windows Server 2012.
Currently the module doesn't explicitly close the file handle. This
wraps the reading of the private key in a try/finally block to ensure
the file is properly closed.
When passing a package version that parses as a number (e.g. `1.9`), the version should be converted to a string before being concatenated to the package name.
add exit_json code to succesfully exit, when you want to delete the already
deleted host.
Without this, playbook fails with
`Specify at least one group for creating host`
which is not correct message.
New module to retrieve facts about existing instance flavors.
By default, facts on all available flavors will be returned.
This can be narrowed by naming a flavor or specifying criteria
about flavor RAM or VCPUs.
- original parameter comment was probably copy&paste error
- new comment highlights that firewall rules can be
added or removed altering this parameter
Session_id is unused in update_session, changed is always specifically
set in all exit_json call, and consul_client.session.destroy return True
or False, and is unused later (nor checked)
TRACE:
while parsing a block mapping
in "<string>", line 33, column 13:
description: resulting state of ...
^
expected <block end>, but found ','
in "lxc_container.RETURN", line 419, column 53:
... "/tmp/test-container-config.tar",
ERROR: RETURN is not valid YAML. Line 419 column 53
- "action" style invoking is a legacy way to call modules
- the examples were updated to the typical style of calling complex
modules:
ovirt:
parameter1: value1
parameter2: value2
...
The os_project module instantiates the openstack cloud object
by passing the module params kwargs.
As the params contain a key named 'domain_id', this is used
for domain in the OpenStack connection, instead of the domain value
the user specifies on the OSCC clouds.yaml or OpenStack envvars.
This fix corrects this by popping the 'domain_id' key, so it we
keep the value but it's not passed later on module.params.
When this module was written back in May 2015 we were using 1.9.x. Being lazy I added to param the objects that the other functions would need. What I have noticed is in 2.0 exit_json is trying to jsonify those complex objects and failing. This PR resolves that issue with the vmware_vm_vss_dvs_migrate module.
@kamsz reported this issue in https://github.com/ansible/ansible-modules-extras/pull/1568
Playbook
```
- name: Migrate VCSA to vDS
local_action:
module: vmware_vm_vss_dvs_migrate
hostname: "{{ mgmt_ip_address }}"
username: "{{ vcsa_user }}"
password: "{{ vcsa_pass }}"
vm_name: "{{ hostname }}"
dvportgroup_name: Management
```
Module Testing
```
ASK [Migrate VCSA to vDS] *****************************************************
task path: /opt/autodeploy/projects/emmet/site_deploy.yml:260
ESTABLISH LOCAL CONNECTION FOR USER: root
localhost EXEC ( umask 22 && mkdir -p "$( echo $HOME/.ansible/tmp/ansible-tmp-1454695546.3-207189190861859 )" && echo "$( echo $HOME/.ansible/tmp/ansible-tmp-1454695546.3-207189190861859 )" )
localhost PUT /tmp/tmpkzD4pF TO /root/.ansible/tmp/ansible-tmp-1454695546.3-207189190861859/vmware_vm_vss_dvs_migrate
localhost EXEC LANG=en_US.UTF-8 LC_ALL=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 /usr/bin/python /root/.ansible/tmp/ansible-tmp-1454695546.3-207189190861859/vmware_vm_vss_dvs_migrate; rm -rf "/root/.ansible/tmp/ansible-tmp-1454695546.3-207189190861859/" > /dev/null 2>&1
changed: [foundation-vcsa -> localhost] => {"changed": true, "invocation": {"module_args": {"dvportgroup_name": "Management", "hostname": "172.27.0.100", "password": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER", "username": "root", "vm_name": "cscvcatmp001"}, "module_name": "vmware_vm_vss_dvs_migrate"}, "result": null}
```
When this module was written back in May 2015 we were using 1.9.x. Being lazy I added to param the objects that the other functions would need. What I have noticed is in 2.0 exit_json is trying to jsonify those complex objects and failing. This PR resolves that issue with the vmware_dvs_portgroup module.
@kamsz reported this issue in https://github.com/ansible/ansible-modules-extras/pull/1568
Playbook
```
- name: Create Management portgroup
local_action:
module: vmware_dvs_portgroup
hostname: "{{ mgmt_ip_address }}"
username: "{{ vcsa_user }}"
password: "{{ vcsa_pass }}"
portgroup_name: Management
switch_name: dvSwitch
vlan_id: "{{ hostvars[groups['foundation_esxi'][0]].mgmt_vlan_id }}"
num_ports: 120
portgroup_type: earlyBinding
state: present
```
Module Testing
```
TASK [Create Management portgroup] *********************************************
task path: /opt/autodeploy/projects/emmet/tasks/deploy/dvs_network.yml:17
ESTABLISH LOCAL CONNECTION FOR USER: root
localhost EXEC ( umask 22 && mkdir -p "$( echo $HOME/.ansible/tmp/ansible-tmp-1454693809.13-142252676354410 )" && echo "$( echo $HOME/.ansible/tmp/ansible-tmp-1454693809.13-142252676354410 )" )
localhost PUT /tmp/tmpeQ8M1U TO /root/.ansible/tmp/ansible-tmp-1454693809.13-142252676354410/vmware_dvs_portgroup
localhost EXEC LANG=en_US.UTF-8 LC_ALL=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 /usr/bin/python /root/.ansible/tmp/ansible-tmp-1454693809.13-142252676354410/vmware_dvs_portgroup; rm -rf "/root/.ansible/tmp/ansible-tmp-1454693809.13-142252676354410/" > /dev/null 2>&1
changed: [foundation-vcsa -> localhost] => {"changed": true, "invocation": {"module_args": {"hostname": "172.27.0.100", "num_ports": 120, "password": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER", "portgroup_name": "Management", "portgroup_type": "earlyBinding", "state": "present", "switch_name": "dvSwitch", "username": "root", "vlan_id": 2700}, "module_name": "vmware_dvs_portgroup"}, "result": "None"}
```
When this module was written back in May 2015 we were using 1.9.x. Being lazy I added to param the objects that the other functions would need. What I have noticed is in 2.0 exit_json is trying to jsonify those complex objects and failing. This PR resolves that issue with the vmware_cluster module.
@kamsz reported this issue in https://github.com/ansible/ansible-modules-extras/pull/1568
Playbook
```
- name: Create Cluster
local_action:
module: vmware_cluster
hostname: "{{ mgmt_ip_address }}"
username: "{{ vcsa_user }}"
password: "{{ vcsa_pass }}"
datacenter_name: "{{ mgmt_vdc }}"
cluster_name: "{{ mgmt_cluster }}"
enable_ha: True
enable_drs: True
enable_vsan: True
```
Module testing
```
TASK [Create Cluster] **********************************************************
task path: /opt/autodeploy/projects/emmet/site_deploy.yml:188
ESTABLISH LOCAL CONNECTION FOR USER: root
localhost EXEC ( umask 22 && mkdir -p "$( echo $HOME/.ansible/tmp/ansible-tmp-1454693788.92-14097560271233 )" && echo "$( echo $HOME/.ansible/tmp/ansible-tmp-1454693788.92-14097560271233 )" )
localhost PUT /tmp/tmpAJfdPb TO /root/.ansible/tmp/ansible-tmp-1454693788.92-14097560271233/vmware_cluster
localhost EXEC LANG=en_US.UTF-8 LC_ALL=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 /usr/bin/python /root/.ansible/tmp/ansible-tmp-1454693788.92-14097560271233/vmware_cluster; rm -rf "/root/.ansible/tmp/ansible-tmp-1454693788.92-14097560271233/" > /dev/null 2>&1
changed: [foundation-vcsa -> localhost] => {"changed": true, "invocation": {"module_args": {"cluster_name": "Foundation", "datacenter_name": "Test-Lab", "enable_drs": true, "enable_ha": true, "enable_vsan": true, "hostname": "172.27.0.100", "password": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER", "state": "present", "username": "root"}, "module_name": "vmware_cluster"}}
```
win_uri uses "Invoke-WebRequest" under the covers, which apparently
uses Internet Explorer to parse a webpage. The problem is if a user
has never run Internet Explorer, it will be unable to do that. The
work around for this is to set the "-UseBasicParsing" flag.
The only advantage to having the Internet Explorer parsed page is
that you can then access the DOM as if it was a powershell
argument. That doesn't seem super useful for Ansible to be able
to do, so I set the default to be "-UseBasicParsing"
When this module was written back in May 2015 we were using 1.9.x. Being lazy I added to param the objects that the other functions would need. What I have noticed is in 2.0 exit_json is trying to jsonify those complex objects and failing. This PR resolves that issue with the vmware_dvswitch module.
@kamsz reported this issue in https://github.com/ansible/ansible-modules-extras/pull/1568
Playbook
```
- name: Create dvswitch
local_action:
module: vmware_dvswitch
hostname: "{{ mgmt_ip_address }}"
username: "{{ vcsa_user }}"
password: "{{ vcsa_pass }}"
datacenter_name: "{{ mgmt_vdc }}"
switch_name: dvSwitch
mtu: 1500
uplink_quantity: 2
discovery_proto: lldp
discovery_operation: both
state: present
```
Module Testing
```
TASK [Create dvswitch] *********************************************************
task path: /opt/autodeploy/projects/emmet/tasks/deploy/dvs_network.yml:3
ESTABLISH LOCAL CONNECTION FOR USER: root
localhost EXEC ( umask 22 && mkdir -p "$( echo $HOME/.ansible/tmp/ansible-tmp-1454693792.01-113207408596014 )" && echo "$( echo $HOME/.ansible/tmp/ansible-tmp-1454693792.01-113207408596014 )" )
localhost PUT /tmp/tmptb3e2c TO /root/.ansible/tmp/ansible-tmp-1454693792.01-113207408596014/vmware_dvswitch
localhost EXEC LANG=en_US.UTF-8 LC_ALL=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8 /usr/bin/python /root/.ansible/tmp/ansible-tmp-1454693792.01-113207408596014/vmware_dvswitch; rm -rf "/root/.ansible/tmp/ansible-tmp-1454693792.01-113207408596014/" > /dev/null 2>&1
changed: [foundation-vcsa -> localhost] => {"changed": true, "invocation": {"module_args": {"datacenter_name": "Test-Lab", "discovery_operation": "both", "discovery_proto": "lldp", "hostname": "172.27.0.100", "mtu": 1500, "password": "VALUE_SPECIFIED_IN_NO_LOG_PARAMETER", "state": "present", "switch_name": "dvSwitch", "uplink_quantity": 2, "username": "root"}, "module_name": "vmware_dvswitch"}, "result": "'vim.dvs.VmwareDistributedVirtualSwitch:dvs-9'"}
```
dnf: name=PACKAGE state=latest is reponsible for two use cases:
- to install a package if not already installed.
- to update the package to the latest if already installed.
The latter use cases is not handled properly as base.upgrade does not
throw dnf.exceptions.MarkingError if a package is not installed.
Setting base.conf.best = True ensures a package is installed or
updated to the latest when calling base.install.
Sign-off: jsilhan@redhat.com
Sign-off: jchaloup@redhat.com
While returning puppet logs as ansible stdout is useful in some cases,
there are also cases where it's more destructive than helpful. For
those, local logging to syslog so that the ansible logging makes sense
is very useful.
This defaults to stdout so that behavior does not change for people.
The bigip_api method was changed in the module_utils function definition
to include the validate_certs option but the bigip_virtual_server module
was not updated accordingly. This patch updates the method so that the
error message below is not returned to the user
received exception: bigip_api() takes exactly 4 arguments (3 given)
This moves the validation of properties to the zfs command itself. The
properties and their choices were not really correct anyway due to
differences between OpenZFS and Solaris/ZFS.
The patch module has a few missing items, and inconsistencies, in its
documentation. A few of which are addressed here.
Within Ansible documentation, the choices for boolean values are
commonly 'yes', and 'no'. We standardise the options on that.
'remote_src' documentation uses 'False' and 'True' for its documentation,
so these have been updated in both the choices and default.
'src' documentation refers to 'remote_src', so is updated to use
the 'no' choice.
'backup' did not describe its options and default at all, so we add
them.
'binary' default used 'False', but specified the type as 'bool' which is
implicitly documented as 'yes'/'no', so we make that 'no' as well.
