Always send websocket messages for
high priority events like playbook_on_stats
Never send websocket messages for
events with no output
unless they are a high priority event type
keep pre-upgrade events in an old table (instead of a partition)
- instead of creating a default partition, keep all events in special
"unpartitioned" tables
- track these tables via distinct proxy=true models
- when generating the queryset for a UnifiedJob's events, look at the
creation date of the job; if it's before the date of the migration,
query on the old unpartitioned table, otherwise use the more modern table
that provides auto-partitioning
Include the EE set on a workflow template in the resolver hierarchy
SUMMARY
This step comes immediately after checking the actual job/template for
an explicitly set EE.
Note that now, because of how jobs are spawned off of workflow nodes,
the call to .resolve_execution_environment() no longer happens in
.create_unified_job(). The job instance within .create_unified_job()
doesn't yet have access to the node that it will be attached to,
making it impossible to use this information in the resolver if called
there.
related #9560
ISSUE TYPE
Feature Pull Request
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Christian Adams <rooftopcellist@gmail.com>
* AWX_ISOLATION_SHOW_PATHS will be shared between containers. Strange
file not found error can crop up when concurrently accessing shared
directories between multiple containers that are bind mounted with big
Z. So make sure we use little z.
This step comes immediately after checking the actual job/template for
an explicitly set EE.
Note that now, because of how jobs are spawned off of workflow nodes,
the call to .resolve_execution_environment() no longer happens in
.create_unified_job(). The job instance within .create_unified_job()
doesn't yet have access to the node that it will be attached to,
making it impossible to use this information in the resolver if called
there.
- base64 encode secret values before creating the secret
- Construct valid .dockerconfigjson
- Cancel jobs where it will obviously fail & error handling
- Check if the secret exists first, then attempts to replace it if it does.
--- Passing this parameter in receptor_params in AWXReceptorJob class
--- Removed AWX_CONTAINER_GROUP_POD_LAUNCH_RETRIES from defaults.py as it was not being used anywhere
--- Removed AWX_CONTAINER_GROUP_POD_LAUNCH_RETRY_DELAY from defaults.py as it was not being used anywhere
Use inventory and env private_data_dir subfolders
SUMMARY
This is another part of trying to get AWX to follow ansible-runner best practices.
See docs:
https://ansible-runner.readthedocs.io/en/stable/intro.html#runner-input-directory-hierarchy
So what runner & people expect is:
.
├── env
│ ├── envvars
│ ├── extravars
│ ├── passwords
│ ├── cmdline
│ ├── settings
│ └── ssh_key
├── inventory
│ └── hosts
└── project
├── test.yml
└── roles
└── testrole
├── defaults
├── handlers
├── meta
├── README.md
├── tasks
├── tests
└── vars
What we were producing was something more like
├── cp
├── env
│ ├── cmdline
│ ├── envvars
│ ├── extravars
│ ├── passwords
│ ├── settings
│ └── ssh_key
├── project
│ ├── ansible.cfg
│ ├── ansible_env.yml
│ ├── async_tasks.yml
<snip>
│ └── vault.yml
└── tmp3evnorsu
There was no inventory folder, and inventory was stored in tmp3evnorsu as a top-level file. Any credential files were also saved in that same structure.
With this change, it's more like:
├── cp
├── env
│ ├── cmdline
│ ├── envvars
│ ├── extravars
│ ├── passwords
│ └── settings
├── inventory
│ └── tmp94xjm0d7
└── project
├── ansible.cfg
├── ansible_env.yml
├── async_tasks.yml
<snip>
└── vault.yml
and any credential files (starting with tmp) are dumped into the env folder.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
AWX VERSION
14.1.0
ADDITIONAL INFORMATION
The motivation for this is that ansible-runner may require specific mounts for every file of directory in the private_data_dir. As such, we really don't want randomized filenames in the top level directory. I don't know for sure this is the direction we are going to go, but this seems like a defensible change on its own, to better adhere to the practices someone familiar with ansible-runner would expect.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Chris Meyers <None>
close db and cache connection in new threads
It's not safe to share or re-use django db connections nor django cache connections across new threads nor processes.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Update with force option if branch_override
Addresses a reported issue where force-pushing to feature branches did not work.
Connect #9371 (comment)
Reviewed-by: Bill Nottingham <None>
