Core Engine Freeze and Module Freeze: 15 August 2017
Core and Curated Module Freeze: 15 August 2017
Community Module Freeze: 29 August 2017
Release Candidate 1 will be 06 September, 2017
Target: Mid-September 2017
- Starting with 2.4, all items that are deprecated will be removed in 4 major releases unless otherwise stated.
- For example: A module that is deprecated in 2.4 will be removed in 2.8
- Ansible will not support Python 2.4 nor 2.5 on the target hosts anymore.
Going forward, Python 2.6+ will be required on targets, as already is the case on the controller.
- Ansible Core Engine and Core modules will be tested on Python 3
- All Core modules now have at least a smoketest integration test.
Additional coverage is welcomed to find more bugs and prevent regressions.
- Communicate with Linux distros to provide Ansible running on Python 3
- Python3 based Ansible packages are now available to run on Fedora Linux
- Proposal found in ansible/proposals issue #35.
- Initial PR of code found in ansible/ansible PR #12797. (done)
- Per plugin configuration (depends on plugin docs below). (WIP)
- New yaml format for config (possibly pushed to future roadmap)
- Extend the ability of the current config system by adding an
ansible-config
command and add the following:
- Dump existing config settings (working, fine tuning)
- Update / write a config entry (pushed to future roadmap)
- Show available options (ini entry, yaml, env var, etc) (working, fine tuning)
(done, needs docs)
- Proposal found in ansible/proposals issue #41.
- Current inventory is overly complex, non modular and mostly still a legacy from inception.
- Configurable list of ‘fact modules’ for
gather_facts
(done)
- Fact gathering policy finer grained (done)
- Make
setup.py
/facts
more pluggable (done)
- Improve testing of
setup.py
/facts.py
(done)
- Namespacing fact variables (via a config option) implemented in ansible/ansible PR #18445. (done)
Proposal found in ansible/proposals issue #17.
(pushed out to future release)
- Over the past couple releases we’ve had some thoughts about how PluginLoader might be better structured
- Load the loaders via an initialization function(), not when importing
the module. (stretch goal, doesn’t impact the CLI)
- Separate duties of
PluginLoader
from PluginFinder
. Most plugins need
both but Modules and Module_utils only need a PluginFinder
- Write different
PluginFinder
subclasses for module_utils and perhaps
Modules. Most Plugin types have a flattened namespace and are single
python files. Modules include code that is not written in python.
Module_utils are vastly different from the other Plugins as they
maintain a hierarchical namespace and are multi-file.
- Potentially split module_utils loader for python from module_utils
loader for powershell. Currently we only support generic module_utils
for python modules. The powershell modules always include a single,
hardcoded powershell module_utils file. If we add generic module_utils
for powershell, we’ll need to decide how to organize the code.
- Pushed to 2.5
- Deprecate (not on standard deprecation cycle)
with_
in favor of loop:
- This
loop:
will take only a list
- Remove complexity from loops, lookups are still available to users
- Less confusing having a static directive vs a one that is dynamic depending on plugins loaded.
- Support for multiple vault passwords. (done)
- Each decrypted item should know which secret to request (done)
- Support requesting credentials (password prompt) as callbacks
- Ability to open and edit file with encrypted vars deencrypted, and encrypt/format on save
(pushed out to future release)
- Make send_callback available to other code that cannot use it.
- Would allow for ‘full formatting’ of output (see JSON callback)
- Fixes static ‘include’ display problem
- Allow plugins to have embedded docs (like modules) (done)
- Update ansible-doc and website to generate docs from these ansible/ansible PR #22796. (ansible-doc working, todo:website)
(done)
- Start using existing group priority variable to sort/merge group vars
- Implementation for this in ansible/ansible PR #22580.
- Documentation of group priority variable
(pushed out to future release)
- Filter on things like “supported_by” in module metadata
- Provide users with an option of “warning, error or allow/ignore”
- Configurable via ansible.cfg and environment variable
- Create import_x for ‘static includes’ (import_task, import_playbook, import_role)
- Any directives are applied to the ‘imported’ tasks
- Create include_x for ‘dynamic includes’ (include_task, include_role)
- Any directives apply to the ‘include’ itself
- New PS/.NET module API (in progress)
- Windows Nano Server support
- Windows module_utils pluginloader (done)
- Refactor duplicated module code into new module_utils files (in progress)
- Evaluate #Requires directives (existing and new: PS version, OS version, etc)
- Improve module debug support/persistence (done)
- Explore official DSC support (done)
- Explore module intermediate output
- Explore Powershell module unit testing (in progress)
- Explore JEA support (stretch)
- Extended become support with network/service/batch logon types
- Module updates
- Split “Windows” category into multiple subs
- Domain user/group management modules (done)
- win_mapped_drive module (done)
- win_hotfix (done)
- win_updates rewrite to require become
- win_package changes required to deprecate win_msi (done)
- win_copy re-write (done)
- Focus on pull requests for various modules
- Triage existing merges for modules
- Module work
- elb-target-groups #19492, #24583. (done)
- alb* #19491, #24584. (done)
- ecs #20618. (in review process)
- Data Pipelines #22878. (in review process)
- VPN #24385. (in review process)
- DirectConnect #26152. (connection module in review process, several more to come)
- Expose endpoint overrides (done)
- Reformat/document module output to collapse internal API structures and surface important data (eg, public IPs, NICs, data disks) (pushed to future)
- Add load balancer module (in progress)
- Add Azure Functions module (in progress)
- Removal of
*_template
modules (done)
- Distributed Continuous Integration Infrastructure (done)
- RPC Connection Plugin (done)
- Module Work
- Declarative intent modules (done)
- OpenVSwitch (done)
- Minimal Viable Platform Agnostic Modules (done)
- All Core and Curated modules will work towards having unit testing. (edit: integration and/or unit tests)
- More bot improvements!
- Bot comments on PRs with details of test failures. (done)
- Test Infrastructure changes
- Shippable + Bot Integration
- Provide verified test results to the bot from Shippable so the bot can comment on PRs with CI failures. (done, compile and sanity tests only)
- Enable the bot to mark PRs with
ci_verified
if all CI failures are verified. (done)
- Windows Server 2016 Integration Tests
- Restore Windows Server 2016 integration tests on Shippable.
- Originally enabled during the 2.3 release cycle, but later disabled due to intermittent WinRM issues.
- Depends on resolution of WinRM connection issues.
- Windows Server Nano Integration Tests (pushed to future roadmap)
- Add support to ansible-core-ci for Windows Server 2016 Nano and enable on Shippable.
- This will use a subset of the existing Windows integration tests.
- Depends on resolution of WinRM connection issues.
- Windows + Python 3 Tests
- Run basic Windows tests using Python 3 as the controller. (partially done, not all planned tests running yet)
- Depends on resolution of WinRM Python 3 issues.
- Cloud Integration Tests
- Run existing cloud integration tests as part of CI for:
- AWS (done)
- Azure (done)
- GCP (pushed to future roadmap)
- Tests to be run only on cloud module (and module_utils) PRs and merges for the relevant cloud provider. (done)
- Test Reliability
- Further improve test reliability to reduce false positives on Shippable. (ongoing)
- This continues work from the 2.3 release cycle.
- Static Code Analysis
- Further expand the scope and coverage of static analysis. (ongoing)
- This continues work from the 2.3 release cycle.