* Providing fix for #47083 in pamd.py
* Providing fix for #47197
* Fixing pep8 errors
* update regex to account for leading dash and VALID_TYPES with dashes as well
* use a results dictionary and clean up unnecessary items
* remove unnessecary return value. action is already reported in invocation output
* make naming consistent across action returns
* fix comparison so it checks equality instead of identity and indentation in update_rule()
* make sure file always has EOF newline
* updated regex to skip spacing between path and args and add rule arg regex to capture complex args
* new module argument parsing code in function and DRY changes
* remove unused has_rule method on PamdService class
* fix error in parse_module_arguments()
* updated args_present action to make it handle key value args and fail on complex bracketed arguments
* pep8 and other fixes so units still work
* suggested change - make version removed 2.8
Co-Authored-By: shepdelacreme <shepdelacreme@users.noreply.github.com>
* add more error proof test to if statement
* Parsing plugin filter may raise TypeError, gracefully handle this exception
and let user know about the syntax error in plugin filter file.
* Test for plugin_filtering
Fixes: #46658
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* ManageIQ module to manage groups
* Fixed typos, more examples on managed_filters.
* Simplified the input if managed_filters
Change return values role_name and tenant_name to role and tenant to macht input params
* Return managed_filters and belongsto_filters instead of filters to better mach the input params
* More generic comparison code from docker_container to docker_common.
* More flexibility if a is None and method is allow_to_present.
Note that this odes not affect docker_container, as there a is never None.
* Update docker_secret and docker_config: simplify labels comparison.
* Added unit tests.
* Use proper subsequence test for allow_more_present for lists.
Note that this does not affect existing code in docker_container, since lists
don't use allow_more_present. Using allow_more_present will only be possible
in Ansible 2.8.
* pep8
* docker_image: Fix up 'changed' event in force mode
This is the same as https://github.com/ansible/ansible/pull/19235 except it applies to all image-building modes (building the image locally, loading the image from an archive, or pulling the image), rather than only when pulling the image.
* Use 'dummy' rather than '_' as unused variable name.
* Add changelog fragment for pull request #33754
Due to refactoring of task_error and wait_for_task method,
SSL thumbprint was lost in error message. This fixes the
retry mechanism of AddHost task.
Fixes: #47563
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* initial commit of facts module for Memset VPS/dedicated server products
* add missing brace
* add integration tests (disabled until we have a mock API to test against)
* bump ansible release version to 2.8
* initial commit of facts module to return usage of a Memstore cloudstorage product
* switch API wrapper to use basic auth instead of passing the api_key in the body
* add integration tests (disabled until we have a mock API to test against)
* bump ansible release version to 2.8
* Controlled params within no section
* Added tests to control params within no section
* Cleaning output_file before creating no-section params and check the content
* addresses comment in PR "s/hate/beverage/g"
* integration test for docker_swarm_service
* ensure stack de-initialization
* Set default value for 'configs' parameter to None
Docker-py uses None as a default value for configs.
Using the same default here allows to create services on older docker
setups (docker_api<1.30).
* Set default value for 'update_order' parameter to None
Docker-py uses None as a default value for update_order.
Using the same default here allows to create services on older docker
setups (docker_api<1.29)
* Set default value for 'publish.mode' parameter to None
Docker-py uses None as a default value for publish_mode.
Using the same default here allows to create services on older docker
setups (docker_api<1.32)
* Allow tests to run on older version of docker.
* remove workarounds for old docker versions
* test correct swarm cleanup
* changelog fragment for docker_swarm_service defaults change
* Add support for POST-as-GET if GET fails with 405.
* Bumping ACME test container version to 1.4. This includes letsencrypt/pebble#162 and letsencrypt/pebble#168.
* Also use POST-as-GET for account data retrival.
This is not yet supported by any ACME server (see letsencrypt/pebble#171),
so we fall back to a regular empty update if a 'malformedRequest' error is
returned.
* Using newest ACME test container image.
Includes letsencrypt/pebble#171 and letsencrypt/pebble#172, which make Pebble behave closer to the current specs.
* Remove workaround for old Pebble version.
* Add changelog entry.
* First try POST-as-GET, then fall back to unauthenticated GET.
* Don't die when get_container is called for container which is terminating during get_container call.
If it terminates between client.containers() and client.inspect_container(),
the module will fail with an error such as
Error inspecting container: 404 Client Error: Not Found ("No such container: xxx")
* Add changelog.
This reverts commit c649d0ea32.
The change results in deadlock in network_cli while it is
waiting to check the return value of recv_ready() which
was added in this commit to improve performance
* Add to k8s_raw docs re: vault-encrypted files
I didn't read the examples far enough and maybe would not have tried all the examples so I ended up creating https://github.com/ansible/ansible/issues/47259 and learned that definition key with lookup works well and we agreed the docs should say something. :)
+label: docsite_pr
* Add note about ansible vault-encrypted files to src: param
* Remove trailing whitespace
* Make changes from feedback
* Make feedback changes
* Fixes#18568
* Commit of the first set of utm modules
* added documentation line for module_utils file
* removed other utm modules for the first pr
* added maintainers to botmeta
* implemented fixes for shippable
* fixed whitespaces and newlines in included doc fragment
* added types and choices to documentation
* fix for E501
* Implemented change requests
* changed utm_utils license to BSD
* changed str() to to_native()
* added a status state that will just return information about my object
* renamed state 'status' to 'info'
* added team_e-spirit to botmeta and added the team as maintainer for the utm_utils
* only return a result if the lookup was not empty. Do not return a null result
* removed info state
* added boilerplate
* made preparation for info-only modules
This PR is fixing following issues:
1) Don't try to check password.
2) Check options.
3) Order wasn't adding at the end, as doc says.
Signed-off-by: Ondra Machacek <omachace@redhat.com>
For some strange reasons the 'swupd' module does not fail if I specify a
non-existing bundle like "xxx". This is inconsistent with other modules, for
example 'dnf'.
The end result is that if someone uses the 'package' module, the result is
inconsistent across OSes: for Fedora trying to install an non-existing module
fails, bot for Clear Linux it just prints a warning and succeeds.
This patch fixes the inconsistent behavior. With this patch 'swupd' will fail
if one tries to install a bundle that does not exist.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
* Avoid misleading PyVmomi error if requests import fails
Requests is imported by the VMware module_utils as an external
dependency; however, because it is in a try/catch block containing the
imports for PyVmomi, if requests fails to import properly, Ansible will
instead complain about PyVmomi not being installed.
By moving the import outside of the try/catch block, if requests fails
to import, an error like the following will be returned:
ImportError: No module named requests
This should result in less confusion.
* catch requests ImportError
* Fixed#47505: Type error in openssl_certificate
* Use to_bytes instead of str.encode in SelfSignedCertificate. Updates #47508
* Use to_bytes instead of str.encode in OwnCACertificate
* Added integration tests for openssl_certificate: selfsigned_not_before/after and ownca_not_before/after
* Add shorthand debugger method to recreate the task from the original ds, and re-template. Fixes#46321
* Fix rst header
* Indicate update_task was added in 2.8
* Major slow down due to using the wrong loader for plugin_types
* traceback due to adding httpapi and cliconf plugins to the plugins we
allow ansible-doc to process
- Added check to see if attribute even exists, if not, it exits.
- Then checks if attribute is already set to value we want to update
it to. If yes, then it exits and changed=False
- Otherwise updates the attribute and changed=True
* Add Scaleway API pagination to server inventory call
* Move Link parsing to helper module
* Correct some PEP8 errors
* Replace AnsibleError with ScalewayException in module_utils since the former doesn't work
* Simplify the regexes to match the intended purpose
* Cleanup helper to conform to review
* Cleanup Scaleway inventory to conform to review
* Flatten the conditional branches structure
* fix a regexp typo
* Add Support of healthcheck in docker_container module
Fixes#33622
Now container can be started with healthcheck enabled
Signed-off-by: Akshay Gaikwad <akgaikwad001@gmail.com>
* Extend docker_container healthcheck (#1)
* Allowing to disable healthcheck.
* Added test for healthcheck.
* Make sure correct types are used.
* Healthcheck needs to be explicitly disabled with test: ['NONE'].
* pep8 fixes
Signed-off-by: Akshay Gaikwad <akgaikwad001@gmail.com>
* Fix bug if healthcheck interval is 1 day or more
`timedelta` object has days too and seconds are up to one day.
Therefore use `total_seconds()` to convert time into seconds.
Signed-off-by: Akshay Gaikwad <akgaikwad001@gmail.com>
* Add test for healthcheck when healthcheck is not specified
This is to avoid the situation when healthcheck is not specified and
treat this as healthcheck is changed or removed.
Signed-off-by: Akshay Gaikwad <akgaikwad001@gmail.com>
* Convert string syntax for healthcheck test to CMD-SHELL
Also add another test case to check idempotency when healthcheck test
is specified as string
Signed-off-by: Akshay Gaikwad <akgaikwad001@gmail.com>
* Playbook fails if minimun docker version is not satisfy for healthcheck
This is to make more consistent with other non-supported options.
Signed-off-by: Akshay Gaikwad <akgaikwad001@gmail.com>
* Add matrix notification module
* try to make ansibot happy
* docs
* fix typo in encoding
* is ansibot happy now?
* change matrix python lib requirement description
* Example formatting & no_log
Thanks for this PR.
Few minor things that are easier for me to just fix, than explain and get you to fix.
* We suggest using `- name:` for examples, as Ansible best practice is to name your tasks
* To prevent secrets being leaked out use `no_log` in argspec
* use `requirements:` in `DOCUMENTATION`
* Clean up argument requirements
* Remove requirements duplicate
* not sure on syntax with these, were adapted from an example elsewhere
* WinRM/PSRP: Ensure shell returns UTF-8 output
This PR makes UTF-8 output work in PSRP shells.
* Add win_command and win_shell integration tests
* Fix tests
* more test fixes
kubeconfig should be loaded if *either* or both of context
or kubeconfig is set (this allows picking a context and default
kubeconfig or picking a kubeconfig with default context)
Fixes#47149
* Behave better if auto_remove and output_logs are combined. Warn if output cannot be retrieved because of auto_remove.
* Add tests.
* Added changelog.
Currently there is no way to reset the custom_compatibility_version to
NULL. If we provide a empty string('') to custom_compatibility_version,
it will fail with error "IndexError: list index out of range" at _get_minor
function.
To reset the custom_compatibility_version, we have to pass None value to
types.Version. The PR fixes the same.
This reverts commit 0e933f76ba.
The tests for this were broken on centos6 because jinja2 does not have
a map filter on that platform. Tests need to be rewritten
This patch fixes an error that occurs when attempting to see if the
netns already exists on the remote device. This change will now execute
`ip netns list` and check if the desired namespace is in the output.
Signed-off-by: Peter Sprygada <psprygada@ansible.com>
* Fix prompt mismatch issue for ios
Fixes#40884#44463
* If the command prompt is matched check if data is
still pending to be read from buffer.
* This fix adds a new timer `buffer_read_timeout`
which will be trigerred after command prompt
is matched and data is attempted to be read from channel.
If not data is present of channel the timer will expire
and response we be returned to calling function.
* Fix unit test failure
* Update to make buffer timeout float
* Update doc and fix review comment
* Fix CI issues
* Update doc
* Fix review comments
* Fix review comments
* Fix AWS EC2 inventory plugin caching of groups
* Added changelog fragment for aws_ec2 caching fix
* Store the AWS query results
The underlying inventory object contains inventory from other sources,
so caching it as ours would be wrong.
It seems easiest and safest to just cache the boto query results
instead.
* Remove unused functions
When a non-existent Gluster volume is to be deleted, i.e.:
gluster_volume:
state: absent
name: vol1-does-not-exist
the "gluster_volume" module would always fail with "volume not found
vol1-does-not-exist". The reason is that the code checked whether the
state is "delete", a value which isn't accepted for the "state"
parameter. Instead the expected value for volume deletion is, as above,
"absent".
Using blockinfile with a custom marker lacking {mark} results in non-idempotency - block is repeatedly inserted on subsequent playbook runs.
+label: docsite_pr
* New module vmware_host_powermgmt_policy
* minor changes
Signed-off-by: Christian Kotte <christian.kotte@gmx.de>
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* New module vmware_host_ipv6
* Minor changes
Signed-off-by: Christian Kotte <christian.kotte@gmx.de>
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
If a "domain" parameter is provided, use it in looking up the user ID.
Additionally, if both a domain and project parameter are provided, then
remove the domain ID from the list of filter criteria after having used
it to look up both the user and the project. OpenStack will not allow us
to apply both a project ID (which implies a domain) and a domain ID as a
search filter.
Fixes#42911
* Add datacenter_name alias
* Improve vcenter_folder module
* add check mode support
* fix VM folder deletion
* fix folder deletion under parent folder
* fix (host) folder creation if two datacenters are present
* Resolve invalid-unary-operand-type.
* Resolve raising-format-tuple.
* Resolve stop-iteration-return.
* Use disable comment instead of fixing logic.
The affected line in _find_address_range will only fail on Python 3.7
and later if the function is called with an empty address list. As an
internal method it is never called in this way, making it a non-issue
for use via public methods.
Using a comment to disable the rule in favor of an ignore.txt entry
since there are no plans to change the logic in the code itself. This
will also prevent any potential future issues being added in other
parts of the code when updating it based on upstream changes.
Fixes: https://github.com/ansible/ansible/issues/45900
This PR fixes the case when the cloud_init_persist was used, but we
still sent use_cloud_init=False, which is in oVirt API evaluated as not
to use cloud_init in first VM execution. This patch is changing it to
send just None, instead of False.
Signed-off-by: Ondra Machacek <omachace@redhat.com>
* Add runtime option to docker_container module
Signed-off-by: Antoine Bardoux <abardoux@nvidia.com>
* Add changelog fragment
Signed-off-by: Antoine Bardoux <abardoux@nvidia.com>
* Add idempotency test for docker_container.runtime
Signed-off-by: Antoine Bardoux <abardoux@nvidia.com>
##### SUMMARY
The azure_rm_cdnprofile module was missing a valid sku 'standard_microsoft'. Added it to the list of valid choices.
##### ISSUE TYPE
- Bugfix Pull Request
##### COMPONENT NAME
azure_rm_cdnprofile
##### ANSIBLE VERSION
```
ansible 2.8.0.dev0
config file = /etc/ansible/ansible.cfg
configured module search path = [u'/home/user/.ansible/plugins/modules', u'/usr/share/ansible/plugins/modules']
ansible python module location = /usr/local/lib/python2.7/dist-packages/ansible
executable location = /usr/local/bin/ansible
python version = 2.7.15rc1 (default, Apr 15 2018, 21:51:34) [GCC 7.3.0]
```
##### ADDITIONAL INFORMATION
It was not possible to deploy using the 'standard_microsoft' sku previously. This addition allows this.
* win_domain: Add missing functionality
* Fix typo: Treshold -> Threshold
* Improvements based on feedback
* moved mode checks to after pre-reqs are installed
The kind: paramter is missing from the original example (which won't actually work). The organization needs to be listed as required (if you don't list it, you will get an error). Also adding additional example using the tower_username and tower_password method.
* reboot: Add timeout in error to help troubleshooting
So we've been hit by 'Timed out waiting for boot_time check' and it was
unclear what timeout was used for the boot_time check. By adding the
timeout value it is easier to understand the used value, and verify if a
change to the timeout is reflected in the output.
* Add changelog fragment
This PR fixes the update check method so it now check also the next_run
configuration of the virtual machine if it exists.
So if previously the VM was updated with new parameters, and then reset
back, the module didn't set the parameters to be set back in next_run.
This PR fixes it so the next run configuration is set back with proper
parameters.
Signed-off-by: Ondra Machacek <omachace@redhat.com>
Bug-Url: https://bugzilla.redhat.com/show_bug.cgi?id=1639894
Signed-off-by: Ondra Machacek <omachace@redhat.com>
* kick off
* done for the day
* beta code and test
* fix a typo
* boto3_conn and boto_exception aren't used in this code, ec2_argument_spec is used but unneeded.
* Returning when find a match avoids doing extra work, especially when pagination is involved
* add new permissions for test
* (output is changed) is preferred over accessing the attribute directly.
* pass the result through camel_dict_to_snake_dict() before returning it.
* AnsibleAWSModule automatically merges the argument_spec.
* deletes the created resources even if a test fails.
* AnsibleAWSModule automatically merges the argument_spec.
* fix typo
* fix pep8
* paginate list_repositories
* specify permissions for test
* cut the unnecessary code.
* add return doc string
* add missed ':'
* fix syntax error: mapping values are not allowed here
* add description for return
* fix syntax error
* rename module name and turn off automated integration test.
* add key_data argument
The key_data argument helps when using a private key stored in Vault.
* Add "version_added" line
Add version_added line to pass automated tests.
* remove unnecessary nesting
* Adding iam_password_policy module
* fixing various issues -- error handling, bugs
* fixing various issues based on tests
* renaming dummy var
* fixing type reference in documentation
* adding int tests and other updates
* removing typo
* fixing auth for int tests
* removing int tests for now
* readding integration tests w/ unsupported designation
* removing conflicting group
* Update aliases
* Fix unused variable
* removed hardcoded loader/plugins list
* updated a few errors to keep orig object
* fix httpapi/cliconf listing
* ansible-doc fixes
* show undocumented as UNDOCUMENTEd
* added missing undoc
* Add support for IOS vlan parsing filter.
Example usage below:
{% set parsed_vlans = vlans | vlan_parser %}
switchport trunk allowed vlan {{ parsed_vlans[0] }}
{% for i in range (1, parsed_vlans | count) %}
switchport trunk allowed vlan add {{ parsed_vlans[i] }}
* Update test_network.py
Add import statement for filter
* Fixed PEP8 issues relating to comments
* Fix PEP8 issues related to blank lines
* Removed magic numbers for line lengths. This should generalize support
to other IOS-like NOS that use similar methods for listing vlans. The
default arguments for line lengths will still be specific to Cisco IOS.
The unit tests for line length are still specific to Cisco IOS.
* An earlier optimization of ansible-doc -l caused failures.
The optimization quickly searches the plugin code for short_description
fields and then uses that in the -l output. The searching was a bit too
naive and ended up pulling out malformed yaml. This caused those
plugins to be omitted from the list of plugins of that type with
a warning that their documentation strings were wrong.
This change makes the documentation parser aware that the documentation
string could have a relative indent for all of its fields which makes it
robust in the face of this particular problem.
* Don't search for space after short_description:
Any whitespace would be valid. In particular newline
1. Add support for SCSI controller creation
2. Support for SCSI Type
3. Warn about Disk Unit Number 7
4. Allow specifying SCSI controller
5. Allow specifying Disk Unit Number
6. Idempotency
7. Remove disks
8. Do not allow to reduce disk
9. More than 15 disks support
10. No SCSI controller and unit number check
11. Thin support
12. Update documentation and example
13. Multiple Datastore and datastore cluster support
14. Check datatype of disk unit number and SCSI controller
15. Handle disk_size when it is int or float
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* Update pydoc
* Update module documentation and example
* Add aliases for name as creation_token
Fixes: #28738
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* Added secure boot enable/disable to vmware_guest_boot_manager. Also added its state to facts
* VMware does not support secure boot when boot_firmware is bios. Add some guardrails to protect the user from misconfiguration
* Address review comments
When multiple virtual machines or templates are found in environment,
module needs folder path. This fix adds facility to specify folder path
for template or virtual machine if multiples are found.
Fixes: #37087
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* Fix in confirmed_commit capability in netconf_config modules
Fixes#46804
* If confirm value is greater than zero or confirm_commit option is set and confirmed-commit
capability is not supported but Netconf server only in that case fail the module
* Update confirm-commit flag
* Minor changes
* Don't execute mount or umount in check mode
* The datastore was mounted or unmounted when you execute the playbook in
check mode.
* Test if running in check mode before mounting or unmounting the
datastore.
* Add support for NFS v4.1 datastores
* Move ansible.compat.tests to test/units/compat/.
* Fix unit test references to ansible.compat.tests.
* Move builtins compat to separate file.
* Fix classification of test/units/compat/ dir.
Python sets the SIGPIPE handler to SIG_IGN. On execv() signal handlers are
reset to their defaults, EXCEPT those that are SIG_IGN which are left ignored.
In Python 3 subprocess.popen explicitly resets the SIGPIPE handler to SIG_DFL,
but unfortunately in Python 2.7 it does not. This leads to subprocesses being
executed with SIGPIPE ignored. This is often a problem with bash scripts which
rely on SIGPIPE to terminate commands in a pipe, but can easily be a problem
with other applications.
This implements the Python 3 behaviour for Python 2.7 by using a preexec_fn.
Adding a value to an attribute or removing a value from an attribute of a LDAP entry throws the exception TypeError: object of type 'filter' has no len()
* Update win_scheduled_task.py to document that the duration of a task trigger can be null, which will cause it to run indefinitely - docs update makes use of suboptions
* Add a fix for the validate-modules schema
* Locate prtdiag even when absent from /usr/bin
On Solaris 8 hosts, this prevents fact collection from aborting with:
Argument 'args' to run_command must be list or string
* Lint fix.
* Style: pass /usr/platform/.../sbin as optional path to get_bin_path().
* Making username password etc non mandatory as we use persistence connection.
Some documentation fine tuning also done.
* To update version which was tested against.
* Adding Bacon switch (NE0152T) to device rules
* Qalthos Review catch
* Updating version
* Changing documentation to add Bacon switch. This is effective since 2.8
* Add url_username and url_password options for url lookup
* Update url.py
Fix documentation
* Fix options documentation
Default value for url_password and url_username is None not an empty string
* Add version added to documentation
* Break command over multiple lines to fix linting
* Fix more linting
* Update version_added in url.py to 2.8
Switches will accept the previously used syntax for /32 ACL entries, but they get changed to host ipaddress in the config file.
This results in the config check failing every time (permit ip 1.1.1.1/32 any log != permit ip host 1.1.1.1 any log), resulting in the ACL getting re-created on every single run of the Playbook.
Making this proposed change to the example will make it work in an idempotent manner.
+label: docsite_pr
This fix initializes current_version to None so that
module proceeds in check mode.
Fixes: #46654
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
If enable_snat is False, this should be used to build the
request, because the default value in the OpenStack Networking
API is True.
Fixes the issue #45915.
* ios_facts: Fix LLDP gathering without neighbors
When LLDP is enabled but no neighbors are present, the following
structure is generated:
...
"ansible_net_neighbors": {
"null": [
{
"host": null,
"port": null
}
]
},
...
If we are not able to find any relevant LLDP information, bail out early
so cases like shown above are handled more gracefully.
* ios_facts: Remove trailing space in lineprotocol
Some Cisco devices (at least CSR1000V 16.6.4) add a space after the
line protocol. This causes the space to be present in the result of
fact gathering:
"ansible_net_interfaces": {
"GigabitEthernet2": {
...
"lineprotocol": "up ",
Be more clear about scanning the output to avoid this behaviour.
* adding explanation of replace_all_instances
How does "replace_all_instances" work? does it increase the number of instances by 1 in the Load balancer and it scales in a server and waits for it to complete its health check, OR does it terminate a node a that causes an autoscale event to happen and then adds a new server?
+label: docsite_pr
* Enable communication to management API over HTTPS.
* Specify version added tags to new parameters
* Set proper parameter type.
* Corrected version_added numbers.
* Extracted commons to ansible utils.
* Fix PEP8 error
* Fix documentation extension syntax.
Fixes#22953
When python is compiled in debug mode, or certain command line flags are
passed, python issues helpful warnings to let users know of files opened
by not closed.
This can be fairly spammy on an ansible run. This change reduces the
number of such warnings by a factor of 10.
* Remove unnecessary workaround
* add test: set RABBITMQ_NODENAME environment variable
The following error occurs:
TASK [rabbitmq_user : Add user] ***
fatal: [testhost]: FAILED! => {
"changed": false,
"cmd": "/usr/sbin/rabbitmqctl -q list_users",
"rc": 69,
"msg": "Error:********@c65c2bc59398'. Please see diagnostics information and suggestions below.\n\nMost common reasons for this are:\n\n * Target node is unreachable (e.g. due to hostname resolution, TCP connection or firewall issues)\n * CLI tool fails to authenticate with the server (e.g. due to CLI tool's Erlang cookie not matching that of the server)\n * Target node is not running\n\nIn addition to the diagnostics info below:\n\n * See the CLI, clustering and networking guides on http://rabbitmq.com/documentation.html to learn more\n * Consult server logs on node test@c65c2bc59398\n\nDIAGNOSTICS\n===========\n\nattempted to contact:********@c65c2bc59398\n * effective user's home directory: /var/lib/rabbitmq\n * Erlang cookie hash: 3MxcYFrJzfhEL+FlUfLlQw==",
"stderr": [...],
"stderr_lines": [
"Error: unable to perform an operation on node 'test@c65c2bc59398'. Please see diagnostics information and suggestions below.", "",
"Most common reasons for this are:", "",
" * Target node is unreachable (e.g. due to hostname resolution, TCP connection or firewall issues)",
" * CLI tool fails to authenticate with the server (e.g. due to CLI tool's Erlang cookie not matching that of the server)",
" * Target node is not running", "",
"In addition to the diagnostics info below:", "",
" * See the CLI, clustering and networking guides on http://rabbitmq.com/documentation.html to learn more",
" * Consult server logs on node test@c65c2bc59398", "",
"DIAGNOSTICS", "===========", "",
"attempted to contact: [test@c65c2bc59398]", "",
"test@c65c2bc59398:",
" * connected to epmd (port 4369) on c65c2bc59398",
" * epmd reports: node 'test' not running at all",
" other nodes on c65c2bc59398: [rabbit]",
" * suggestion: start the node", "",
"Current node details:",
" * node name: rabbitmqcli2@c65c2bc59398",
" * effective user's home directory: /var/lib/rabbitmq",
" * Erlang cookie hash: 3MxcYFrJzfhEL+FlUfLlQw==", ""
],
"stdout": "",
"stdout_lines": []
}
* node parameter: fix default value
'rabbit' is the default value mentioned in the module documentation.
This fix checks if executable exists in local connection specified using
ANSIBLE_EXECUTABLE variable.
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
* module_utils.urls: add fetch_file function
* apt: use fetch_file instead of own download()
* unarchive: use fetch_file instead of own codecopy
* apt: add test for deb=http://…
* unarchive: add test for a remote file download and unarchive
* yum: replace fetch_rpm_from_url by fetch_file
* use NamedTemporaryFile
* don't add a dot to fileext, it's already there
* allow update_cache as stand alone operation for yum/dnf
Fixes#40068
Signed-off-by: Adam Miller <admiller@redhat.com>
* make sanity tests happy
Signed-off-by: Adam Miller <admiller@redhat.com>
On connection errors the influxdb python module raises an exception with
an empty message leading to a fail_json with an empty error message:
265d14736b/influxdb/influxdb08/client.py (L252)
Fail with a more descriptive error in this case.
Currently, Ansible interprets variables with a True|False value as
boolean. This causes the vmware_deploy_ovf module to break, because
it can only accept string values as properties.
This fix checks if a value is boolean, and converts it to a string
if it is. Since integers do not seem to be causing the same error, this
is the only check we appear to need.
After completion, OVF properties that are boolean can be specified as
yes|no or true|false.
Closes: #45528
* Add chdir support to win_package
* Add chdir support to uninstall too.
* Update docs to account for uninstall.
* Correct variable substitution for hashtable.
* added changelog fragment
* Added module to manage manageiq tenants
* Fixed pep8 and module validation issues for manageiq_tenant
* Changes based on review
- Added parent attribute to specify the parent
- Cleanup return of tenant
- Update documentation on quotas
* Updated version added to 2.8
* Allow to specify images by hash for docker_container and docker_image_facts.
* flake8
* More sanity checks.
* Added changelog.
* Added test.
* Make compatible with Python < 3.4.
* Remove out-commented imports.
* Don't simply ignore container in present() if image is not specified.
* Use image from existing container for recreation if not specified.
* Added changelog.
* Improve comment.
* Ovirt_disk - Add ability to refresh passthrough disks
A new paramter "host" is added. When this paramter is specified the newly
created disk or an existing disk will refresh its information about the
underlying storage via the specified host. This oprion is only valid for
passthrough disks. This option requires at least the logical_unit.id to be
specified.
* Host search modification and some other minor changes.
* pep8 correction
* ovirt add snapshot delete after x days
* update logic of deleting snapshot and rename it
* update docs and names of vars
* update docs syntax
* Update ovirt_snapshot.py
* put remove logic to function
* use default removing function for snapshot
* win_reboot: Fix for ignore post_reboot_delay
This fixes an issue where win_reboot would be ignoring the provided
post_reboot_delay (and on Windows timing/waiting is everything!)
This must be backported to the v2.7 branch.
* Merge post-reboot handling into run()
* Adding a basic get lookup for rabbitmq.
* Always return a list
* If content type is JSON, make accessible via dict.
* Fixed incorrect json.loads variable and missing raise
* Change to document returned data
* Fixed pep8 issues
* Adding integration testing
* Moving lookup intgration tests to new target
* New rabbitmq lookup plugin (#44070).
* New rabbitmq lookup plugin (#44070).
* PR review feedback updates
* Testing pika is installed
* Minor mods to tests
* Check if connection is already closed or closing
* Updated tests and connection testing
* PR review feedback updates
* PR review include ValueError in AnsibleError output
* Suggesting to use set_fact when using returned variable more than once.
* Cleaned up some tests, added some notes and handling connection closure on some exceptions.
* Removed finally statement and added some additional error handling.
* Added some additional error handling.
* PR review updates.
* Additional integration tests and removing return in finally
* Updated version
* Changing back to running tests on ubuntu.
* Additional tests
* Running tests on Ubuntu only
* Fixing syntax error
* Fixing ingtegration tests and a string/byte issue
* Removed non-required test and fixed BOTMETA
* Trying to fix integration test failure on ubuntu1404
* Some issues occured when handling messages from the queue with to_native. Switching to to_text resolved the issues.
* Renaming channel to queue (thanks dch). Disabling trusty tests.
* macports: Replace update_ports with selfupdate
- Macports discourages use of `port sync` and recommends using `port
selfupdate` instead.
- Keep `update_cache` and `update_ports` as aliases.
- No longer require the `name` parameter so that `selfupdate` can be
used in a task by itself.
* macports: Add upgrade parameter
- New upgrade parameter which can be used to upgrade all outdated ports.
* Add changelog fragment
Fixes#35120 : the redis cache plugin keeps key/value
entries in an in-memory cache to avoid hitting the
redis database each time.
The problem is that a cache entry is only set when
a value is get or set but it is always deleted when
trying to delete a value.
When the --flush-cache ansible-playbook option is used,
the redis cache plugin is first asked to remove every
entry corresponding to every hostname present in the inventory.
As no value as been set/get so far, it then tries to delete
an unexisting value from the cache and hence crashes with
a KeyError exception.
* win_exec: refactor PS exec runner
* more changes for PSCore compatibility
* made some changes based on the recent review
* split up module exec scripts for smaller payload
* removed C# module support to focus on just error msg improvement
* cleaned up c# test classifier code
* New option JUNIT_TASK_RELATIVE_PATH to output relative paths.
It makes the output in Jenkins much more readable as the absolute
path was taking the whole width of the screen, leaving no space for
the actual message.
A makedirs instead of mkdir slipped in as to make the creation of
the output directory recursive, hope this is OK.
* Add version_added field to doc set to 2.8
* Separate networking tools that may be used by modules outside of networking so changes to networking-only utilities don't trigger AWS integration tests
* Add unit tests for moved network utils
* Add comment to prevent imports from being mistakenly removed
* Move to_bits as well
* Check that union Jinja filter can be chained
* set filters: fix unexpected templating type error
this error occurs with Jinja 2.10 since 32ec69d827,
for example when union filters are chained:
$ ansible all -i localhost, -mdebug -a"msg={{ []|union([])|union([]) }}"
localhost | FAILED! => {
"msg": "Unexpected templating type error occurred on ({{ []|union([])|union([]) }}):
unsupported operand type(s) for +: 'set' and 'list'"
}
* fix callback Plugin json to support global stat set by set_stats module
* refactor to not break compatiblity, add trailing comma on output dict
* Remove sorted, as it not needed
* refactor, to sivels better code
* clean some code messup
* add changelog fragment
* added missing new line
* fix pep8 stuff
In a scenario where you've set your logstash heap size high, you may need to set your heap lower in order to execute logstash-plugin while logstash is still running.
Here's an example of someone running into insufficient memory while executing logstash-plugin: https://discuss.elastic.co/t/logstash-plugin-and-jvm-options/99646
This scenario is probably fairly common, so it may be worth documenting as an example.
+label: docsite_pr
* docker_container: Honour stop_timeout when creating docker containers (#43814)
* Adjusting description to what actually happens.
See docker-py changelog for 2.7.0: 'APIClient.stop will
no longer override the stop_timeout value present in the
container’s configuration.'
* Add a test whether stop_timeout can be configured for the container.
* Added changelog.
* Integrate with comparisons (by default, ignore stop_timeout value for restarts; will be configurable with PR ansible/ansible#44789).
* Fix config change code and tests (#2)
* Improving wildcard test.
* Using correct config.
Previously, when the active SELinux value was the same as the desired value, regardless of the value of `persistent`, the module would simply exit successfully, assuming no need for update. This made the assumption that the active and persistent values should be the same, but that is not always the case. This modification treats both the active and persistent values separately when checking for the need to update. Note that the persistent update mode, however, will still update the active as well as the persistent value. If this is not desired, it is possible to separately toggle the active value alone.
Other changes:
* Make the check mode actually perform checks for changes
* Organizes semanage commands into set of logical steps
This prevents a stack trace in Python 3 when the result is an empty file since
the file is open in binary mode and a native string in Python 3 is str,
not bytes.
The stdout and stderr values returned from self._low_level_execute() are text, not bytes. This results in an error in Python 3 since str and bytes cannot be concatenated.
Changing to unicode type allows this to work without error on Python 2 and Python 3.
* modules/systemd: fix logic: allow scope to default to 'system'
Fix logic introduced in 7ea909418e: if 'scope' param is not specified,
it defaults to system, but the value of module.params['scope'] is None,
not 'system' - so allow for that.
* modules/systemd: fix logic: disabled means disabled
Fix logic determining whether a service with both systemd and initd files is enabled or disabled.
In situations where systemd thinks service is disabled, but rc.d symlinks mark it as enabled,
this module wrongly assumes the service is enabled.
Fix this logic: disabled means disabled
Only when the output from does NOT include disabled, consider the status of rc.d symlinks.
This essentially replicates the fixes done to the systemd handling in the "service" module in 3c89a21e0cFixes#22303Fixes#44409
* Fix spelling of 'separate' throughout.
* Various cleanups in the User Guide for Vault.
- Fix spelling of 'algorithm'
- Fix indentation of nested list in payload format
- Fix mysterious refernce to 'b_pkey1'.
- Fix reference to newline as '\n': the backslash is lost when rendered
to the docs website. Specify the hex value for newline instead of the
backslash escape.
* Fix formatting
* Update vault.rst
A lot of modules have a short_description with a trailing dot even
though we don't want trailing dots in the index. This change removes
it when creating the document index.
* Fail on attrs that don't work for includes
* Prefix with VALID_ to make it more clear
* Sort keywords
* Remove unnecessary list
* Adjust dynamic/static checking
In DNF < 3.0 are lists, and modifying them works
In DNF >= 3.0 < 3.6 are lists, but modifying them doesn't work
In DNF >= 3.6 have been turned into tuples, to communicate that
modifying them doesn't work
Further explanation of this is available via Adam Williamson from
the Fedora QA Team.
https://www.happyassassin.net/2018/06/27/adams-debugging-adventures-the-immutable-mutable-object/
Signed-off-by: Adam Miller <admiller@redhat.com>
* win_nssm: add failing tests for issue #44079
* win_nssm: use Run-Command instead of Invoke-Expression to prevent interpretation issue
Fix#44079
* win_nssm: add more failing tests
These tests highlight several issues with this module:
* Service not started when state=started
* Errors with app_parameters (see #25265)
* Exception when passing several dependencies separated by comma as specified in doc
* win_nssm: fix service not started when state=started
Nssm status returns a multiline output that doesn't match any of the strict patterns in the switch statement.
* win_nssm: fix incorrect separator in doc for service dependencies
The dependencies parameter works with space as separator, but not with comma as shown in the documentation
* win_nssm: fix error with app_parameters parameter
Fix#25265
* win_nssm: add idempotence tests
* win_nssm: fix several idempotence issues and misbehaviors
Add missing space between arguments when app_parameters contains several keys.
Use Argv-ToString and Escape-Argument to improve arguments handling (parameters with quotes, backslashes or spaces).
* win_nssm: test parameters with spaces, quotes or backslashes
* win_nssm: restore comma as separator for service dependencies
Revert commit ddd4b4b
* win_nssm: restore support of string as dict form for app_parameters and remove support of literal YAML dict
* win_nssm: wrong variable in tests
* openshift inventory: fix exception when auth fails
Fix 'ForbiddenError' object has no attribute 'message':
[WARNING]: * Failed to parse test.yml with openshift plugin: 'ForbiddenError' object has no attribute 'message'
File "ansible/lib/ansible/inventory/manager.py", line 270, in parse_source
plugin.parse(self._inventory, self._loader, source, cache=cache)
File "ansible/lib/ansible/plugins/inventory/openshift.py", line 122, in parse
self.setup(config_data, cache, cache_key)
File "ansible/lib/ansible/module_utils/k8s/inventory.py", line 58, in setup
self.fetch_objects(connections)
File "ansible/lib/ansible/module_utils/k8s/inventory.py", line 250, in fetch_objects
super(OpenShiftInventoryHelper, self).fetch_objects(connections)
File "ansible/lib/ansible/module_utils/k8s/inventory.py", line 81, in fetch_objects
namespaces = self.get_available_namespaces(client)
File "ansible/lib/ansible/module_utils/k8s/inventory.py", line 95, in get_available_namespaces
raise K8sInventoryException('Error fetching Namespace list: {0}'.format(exc.message))
Don't try to get 'message' attribute from:
- K8sInventoryException instances
- Exception instances
- KubernetesException instances (because KubernetesException can be
Exception)
* move k8s/OpenShift inventory plugin dedicated code
inventory plugin specific code should not be located in
lib/ansible/module_utils directory. Then ansible.utils methods can be
reused (for example Display).
* Remove unused class variables 'helper'
unused since 4d77878654.
* Update win_robocopy.py
<!--- Show example of UNC path copy of a single file, noting trailing backslash required -->
+label: docsite_pr
* addresses jborean comment, removes double quotes
* Remove trailing slash part
* Added feature facts to nxos_facts
* Fixed ansibot indentation
* Resolved ansibot whitespace missing after ','
* Per PR suggestion, created method in base class to gather switch config
and store in global variable to prevent multiple calls to switch.
* Addressed ansibot blank line & whitespace after ,
* Enable support for launch_templates in ec2_asg
* Fix asg create with LT and no version number
* Update mutually exclusive list
* Better function names
Maintain one Templar for the lifetime of VariableManager, calling
set_available_variables() prior to each use, enabling _get_filter()'s
cache to function correctly.
It does not seem possible for concurrent calls into one (non-copied)
VariableManager instance, and so it need not be reentrant. If that
became a requirement, serializing its or Templar's entry points would be
fine, as it's so CPU-heavy other threads will only fight with it for the
GIL anyway.
Reduces _get_filters() runtime 91%, get_vars() runtime 19%, function
call count 16%, overall runtime 10%.
Tested aginst dummy load comprised of the 12 disabled steps of
debops.auth with an inventory of 80 hosts, which stresses variable
processing and task setup. Before:
7447296 function calls (7253994 primitive calls) in 32.611 seconds
Ordered by: cumulative time
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.000 0.000 32.762 32.762 ansible-playbook:3(<module>)
1 0.007 0.007 31.733 31.733 ansible-playbook:21(<module>)
...
1371/971 0.671 0.000 21.332 0.022 manager.py:154(get_vars)
...
3044 0.315 0.000 5.166 0.002 __init__.py:295(_get_filters)
After:
6252978 function calls (6059638 primitive calls) in 29.055 seconds
Ordered by: cumulative time
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.000 0.000 29.218 29.218 ansible-playbook:3(<module>)
1 0.007 0.007 28.159 28.159 ansible-playbook:21(<module>)
...
1371/971 0.675 0.000 17.211 0.018 manager.py:154(get_vars)
...
3044 0.028 0.000 0.441 0.000 __init__.py:295(_get_filters)
* allow jinja2 unique filter compat
* detect if unique is provided, fallback with warning
* handle j2 specific params
* now all filters using unique must pass environment
* added env to tests
also normalized on how we normally import and use exceptoins
* Start of work on pylint plugin to catch due/past-due deprecated calls
* Improve deprecated pylint plugin
* Catch call to AnsibleModule.deprecate also
* Skip splatted kwargs, we can't infer that info
* Add error for invalid version in deprecation
* Skip version if it's a reference to a var
* Disable ansible-deprecated-no-version for displaying deprecated module info
* fix comments
* is None
* Force specifying a version, this can be disabled on a per case basis
* Disable ansible-deprecated-version by default
* Remove to look for 2.8 deprecated
* Revert "Remove to look for 2.8 deprecated"
This reverts commit 4e84034fd1.
* Add script and template used for creating issues for deprecated issues
* Fix underscore var
* Update synchronize.py
If you want a different rsync binary on the local side, you need to set task variable ansible_rsync_path.
See examples.
Variable ansible_rsync_path looks to not be documented anywhere. If documented, needs to be said that is does not belong to synchronise options, instead belongs to tasks. (Sorry, I have no better wording)
* [rabbitmq_binding] Fix the quoting of vhost and other names, which was broken in PR #35651
* Merge missing urllib_parse.quote from PR #42422
* Missed one line, where also needs to be escaped, i.e., the destination
* Various changes to nxos_logging. Plus added purge capibility.
* Made a few new nxapi_logging test cases conditional based on version
and/or platform.
* Addressed PR comments and ansibot shippable. Fixed up nxos_logging documentation format.
* Addressed ansibot shippable issues with whitespaces and documentation.
* Resolved ansibot codestyle trailing whitespace
* adds support for null values to the ternary filter
This change adds a third optional argument to the ternary filter to
handle a null value. If the third option is specified null and false
are treated differently.
For instance, take the following example:
{{ enabled | ternary('no shutdown', 'shutdown') }}
If enabled == True, then 'no shutdown' is used.
If enabled in (False, None), then 'shutdown' is used.
With this change the following is possible:
{{ enabled | ternary('no shutdown', 'shutdown', omit) }}
If enabled == True, then 'no shutdown'
If enabled == False, then 'shutdown'
If enabled == None, then omit
* update documentation with example of filter
* update filter documentation example per comments
* fix logic error in user_guide example
* Ensure that the value of PLUGIN_FILTERS_CFG is treated as type=path, and that we use the standard section of 'defaults' instead of 'default'
* deprecate the default section
* Don't add version_added for the corrected section
<!--- Your description here -->
The documentation states the use of the "value" attribute for environment variables while this should also be the "job" attribute.
+label: docsite_pr
* Sorting args.
* Doing comparisons of options with container parameters in a more context-sensitive way.
This prevents unnecessary restarts, or missing restarts (f.ex. if parameters are removed from ``cmd``).
* Make blkio_weight work.
* Fix cap_drop idempotency problem.
* Making groups idempotent if it contains integers.
* Make cpuset_mems work.
* Make dns_opts work.
* Fixing log_opts: docker expects string values, returns error for integer.
* Adding tests from felixfontein/ansible-docker_container-test#2.
* Make uts work.
* Adding changelog entry.
* Forgot option security_opts.
* Fixing typo.
* Explain strict set(dict) comparison a bit more.
* Improving idempotency tests.
* Making dns_servers a list, since the ordering is relevant.
* Making dns_search_domains a list, since the ordering is relevant.
* Improving dns_search_domains/dns_servers.
* Fixing entrypoint test.
* Making sure options are only supported for correct docker-py versions.
* Update yaml-style in password-lookup example
##### SUMMARY
Update the yaml-style in a password-lookup example to match best-practices.
##### ISSUE TYPE
- Docs Pull Request
##### COMPONENT NAME
password_lookup plugin
##### ANSIBLE VERSION
devel
* remove whitespace
* fixes#45941
* corrects regression introduced by #26104; when the resource group doesn't exist, the module exits prematurely with an error instead of creating it.
* Unify login behavior between 1Password lookup plugins and module
- Use the same names for all credential aspects
- Only require the minimal amount of information for each
- Add more examples
* Change parameter terms
- use terms in line with 1Password documentation.
- update examples
- update tests
* Improve error messages in lookup plugin
* Unify onepassword_facts with lookup plugins
- use same methods and logic for signing in or reusing existing session
- unify terms with lookup plugins
* Change rc test for determing login
An rc other than 1 can be returned when a current login session does not exist.
* Create AnsibleModuleError class
ansible.errors is not available to modules, so create an AnsibleModuleError class within the module
Do not user os.path.expanduser since this is already done by virtue of the type being "path" in the argument spec.
* Add note about risk with fact caching sensitive data
* Add note on op version that was used for testing
This is because underscores are illegals in hostnames and users might
want to create the same host names are the name of the Linode machine
(as in the report in #30059).
Closes https://github.com/ansible/ansible/issues/30059.
* Update azure_rm_virtualmachine.py
A lot of Azure images are not cloud-init ready and need to be prepared manually before attempting to use the custom_data option of this module. Adding a line to the description to make others aware. If, like me, they are used to working with AWS AMIs that all seem to have cloud-init baked in, this could prevent some troubleshooting as to why their custom_data scripts aren't running in Azure.
* fixing addition to azure_rm_virtualmachine.py
* Update azure_rm_virtualmachine.py
* final update
* fixed url syntax
* nuked trailing whitespace
* Ensure that the src file contents is converted to unicode in diff info. Fixes#45717
* Fix up and cleanup
* The diff functionality in the callback plugins should have the
to_text() calls removed since we're now doing it in ActionBase
* catching of UnicodeError and warnings in the callback diff
functionality from 61d01f549f haven't been
needed since we switched to to_text so remove them.
* Add a note to ActionBase's diff function giving an example of when the
diff function will be inaccurate and how to fix it
* Fix callback get_diff() tests
I believe the unittests of callback's get_diff() were wrong. They were
sending in a list where strings were expected. Because previous code
was transforming the lists into strings via their repr, the previous
tests did not fail but they would have formatted the test cases output
in an odd way if we had looked at it.
* Try to intuit proper plugins to send to ansible-connection
* Move sub-plugins to init so that vars will be populated in executor
* Fix connection unit tests
* win async: use async_dir for the async results file directory
* tried to unify POSIX and PowerShell async implementations of async_dir
* fix sanity issue
* handle yum and dnf lockfiles - fixes#44120
Signed-off-by: Adam Miller <admiller@redhat.com>
* fix logic problem to properly check for dnf lockfile glob
Signed-off-by: Adam Miller <admiller@redhat.com>
* Added documentation for collectstatic --clear.
Added a description of the --clear argument used with the Django management
command, collectstatic. (When #1810 added this feature the documentation
was not updated).
no such option "config" for diff_against argument.
Further below, there is an example of eos_config using diff_against with a valid option.
+label: docsite_pr
Also fixes an error related to the required parameters. The l3out parameter is required but only 'name' was being accepted, and that should only be the alias.
* efs.py: Add support for EFS provisioned throughput
* efs_facts.py: Add support for EFS provisioned throughput
* efs_facts integration tests updated with provision throughput
* efs_facts: Tests refactoring - add failure and success playbook according to botocore version.
* efs_facts: More tests and new option descriptions adjustment
* efs_facts tests renamed to efs
* Improve iam_group exception handling
Use AnsibleAWSModule for iam_group and handle BotoCoreErrors
as well as ClientErrors. Use fail_json_aws to improve error messages
* Add minimal iam_group test suite
Update some of the read-only IAM permissions (this is not sufficient
to run the test suite but it gets further than it did until it tries
to add a (non-existent) user)
* Clean up after tests
* Add the key_name/value_name options to dict2items - as with items2dict, allow users to configure the key/value name for dict2items, add "version added" and examples
* Added example to register result of a ping test, including a wait_for command.
* added names to tasks
* updated examples to remove deprecated commands etc
* removed group_var
Since 'organization' is the spelling used accross all others modules,
I think it would be better to at least have this one as a alias.
Organisation is the UK/Australia/NZ spelling, while organization is the
US one.
* Link to the Galaxy platforms list from the meta file template
Fixes https://github.com/ansible/galaxy/issues/52.
* make ansible/ansible text match mazer text on platforms
