No result_traceback is blank, not null
I kept staring at a failure where:
cannot create resource
was expected in result_traceback, but it was blank. The job had "error" status, sure enough.
Then I re-created, sure enough, the job was right there - error status, no explanation, no traceback, no indication of what happened. We did have logs though (which I still can't fully disambiguate).
After that initial confusion, I realized this is a server-side bug. The test is happy with what happens after this change.
I can re-organize this diff if needed.
Reviewed-by: Elijah DeLee <kdelee@redhat.com>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Reviewed-by: Shane McDonald <me@shanemcd.com>
bumped django version to 2.2.20 in devel
SUMMARY
Bumped Django version from 2.2.16 -> 2.2.20 for the devel environment. All builds, tests and tests consuming the AWX_USE_FIPS flag passed in Jenkins.
AWX Issue - ansible/tower#4930
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.2.2
Reviewed-by: Alan Rominger <arominge@redhat.com>
Fix pylint 2.9 errors in awx collection
SUMMARY
fixes errors like these
ERROR: plugins/modules/workflow_launch.py:131:4: consider-using-dict-items: Consider iterating with .items()
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.2.2
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Make quay.io the registry URL default
SUMMARY
It's unclear what format the registry "Authentication URL" should be.
It's too late to change the strings because of translations, ping @rooftopcellist, but this change would clarify the format without adding wording.
ISSUE TYPE
Feature Pull Request
Docs Pull Request
COMPONENT NAME
API
ADDITIONAL INFORMATION
Seems to behave pretty good in the UI.
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
reload settings on fork or new thread
Reloading settings will re-import cache. Django cache is thread aware,
but keeping an old reference to the cache avoids the awareness.
The performance team that was hitting this every time before. With this change they are unable to recreated.
Reviewed-by: Shane McDonald <me@shanemcd.com>
Close file before returning
Resolves
/home/alancoding/repos/tower-qa/tests/lib/plugins/pytest_restqa/plugin.py:122: ResourceWarning: unclosed file <_io.TextIOWrapper name='/home/alancoding/repos/tower-qa/scripts/resource_loading/data_latest_loading.yml' mode='r' encoding='UTF-8'>
qe_config.resources = PseudoNamespace(yaml_file.load_file(config.option.resource_file))
We have this same pattern earlier in the file.
Reviewed-by: Shane McDonald <me@shanemcd.com>
Fix awx collections copy_item warning keyword error
SUMMARY
warn() does not take keyword 'msg'
add testing around the copy_from parameter for project creation
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.2.1
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Remove more freenode from docs
SUMMARY
Removing more mention of Freenode, since we moved to Libera.
ISSUE TYPE
Docs Pull Request
Reviewed-by: Shane McDonald <me@shanemcd.com>
Update Dockerfile.j2
SUMMARY
Jobs unable to start because podman trying to use systemd cgroup manager. See error below :
WARN[0000] Failed to add conmon to systemd sandbox cgroup: dial unix /run/systemd/private: connect: no such file or directory
Error: OCI runtime error: systemd cgroup flag passed, but systemd support for managing cgroups is not available
related #10099
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.0.0
ADDITIONAL INFORMATION
According to this PR containers/podman#7009, podman switch references from libpod.conf to containers.conf.
According to containers.conf man (https://github.com/containers/common/blob/main/docs/containers.conf.5.md), configuration file is a TOML file but engine section declaration is missing.
thanks to @Siorde too 👍
Reviewed-by: Chris Meyers <None>
Reviewed-by: Shane McDonald <me@shanemcd.com>
List platforms that developers regularly run the dev env on
SUMMARY
Other platforms are generally untested.
ISSUE TYPE
Docs Pull Request
Reviewed-by: Julen Landa Alustiza <None>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Update and expand docs/ folder for EEs
This has some new content about EE precedence, which I don't think we've documented anywhere else, thinking of @tvo318 here, content was developed by @jbradberry
(I think the numbers 2 and 3 in the global job default EE may still be subject to revision, just a warning)
@shanemcd #10324 is incorporated into this
This mentions @rebeccahhh's venv migration stuff, but I'm trying to write the absolute minimum possible while still mentioning migration.
Reviewed-by: Rebeccah Hunter <rhunter@redhat.com>
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Jeff Bradberry <None>
Handle inventory types where Automation Hub collection names differ
Some collections will be moving, which is #10323
Yet, other collections will change name, which this is intended to address.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Sarabraj Singh <singh.sarabraj@gmail.com>
Reviewed-by: Elijah DeLee <kdelee@redhat.com>
Remove isolated instances and groups
SUMMARY
before we remove the fields that allow us to correctly identify them.
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Chris Meyers <None>
Add validation for EE image field name
related #10431
SUMMARY
Adds validation for the EE image name
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.2.1
Reviewed-by: Seth Foster <None>
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Reviewed-by: Rebeccah Hunter <rhunter@redhat.com>
Reviewed-by: Jeff Bradberry <None>
introduced a pre-flight check for postgres 12
SUMMARY
Introduced a pre-flight check to check the version of Postgres 12 before the application starts - ansible/tower#4940
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.2.0
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Sarabraj Singh <singh.sarabraj@gmail.com>
Reviewed-by: Christian Adams <rooftopcellist@gmail.com>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Prefer installer defaults over user-defined global EEs
The installer knows better than the user
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Reviewed-by: Elijah DeLee <kdelee@redhat.com>
Add Thycotic DevOps Secrets Vault support
SUMMARY
Add a Credential Plugin for Thycotic DevOps Secrets Vault
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
UI
AWX VERSION
awx: 19.1.0
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Adam C. Migus <adam@migus.org>
Reviewed-by: Matthew Jones <bsdmatburt@gmail.com>
Make ExecutionEnvironment awxkit class able to handle null values in …
This is how the Project class does it, and I can confirm it makes factories.execution_environment(organization=None) work correctly.
Reviewed-by: Jeff Bradberry <None>
Fixes bug where source list page would crash if first sync was running
SUMMARY
link #10411
The ID being passed to the cancel button was incorrect as well. We want to reference current_job not last_job.
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Kersom <None>
Reviewed-by: Sarah Akus <sarah.akus@gmail.com>
Addresses bug where advanced search by groups wasn't working on host list
SUMMARY
link #10400
OK so this got a bit more complicated than I wanted it to but I think on the whole this is a nice improvement. In the initial bug report it was noted that we were ignoring the or/not operators when searching on a related key (in this case it was groups). We were also ignoring the modifier at the end (icontains, exact, startswith, etc). Since we were always defaulting to __search for these types of keys this was actually valid. The API doesn't like us attempting to do something like ?or__groups__search=foo or ?groups__search__icontains because they aren't valid. So, I changed the UX a little bit. A user can still do a fuzzy search on a related key but the prefix is disabled in this case:
I changed the third dropdown (specifically for related keys) to contain the following options: search, id, name__icontains which I think would be the three most general cases. id and name__icontains do allow for prefixing (or, not, and).
This should make searching on related keys a little bit easier and a little bit more flexible since name and id weren't possible in the UI before.
Once place where this is a little different is the host filter in the smart inventory form. Using __search currently throws an error when the user attempts to save. I believe this is an API bug and will file it as such but for now I prevent users from attempting to use __search by removing it as an option for this particular list. When this bug gets fixed in the API we can remove this logic.
I also noticed that there was a bug where or__ search terms were being converted to and when using the Smart Inventory button on the host list so I fixed that.
Here's the fix in action:
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
fixes next button and internationalizes nav buttons on wizard
SUMMARY
Resolves#10404 and it internationalizes the next button.
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
AWX VERSION
ADDITIONAL INFORMATION
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Pull brand name from default.strings.json
SUMMARY
We were pulling this string from variables.js.
There were several places where the brand name was not being pulled dynamically - this PR addresses that. This changes the way we set the document title a little bit. Since we needed access to a dynamic string, I went down the JS route to set this.
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Sarah Akus <sarah.akus@gmail.com>
Changes settings revert all to send DELETE on individual endpoint rather than PATCHing
SUMMARY
Resolves#3831#3831 (comment) is a good explanation of the state of the new UI as well as the old. The Misc System settings were the most problematic part of this since that form pulled fields from both /api/v2/settings/system and /api/v2/settings/authentication. In order to revert these fields with a DELETE request I needed to split this form up into Miscellaneous System and Miscellaneous Authentication. The Activity Stream Settings also needed to be absorbed by the Miscellaneous System Settings.
All settings forms (except noted below) will now issue a DELETE request to revert all the fields for a particular settings subsection.
There was one form that I did not change the Revert All functionality on and that was the LDAP Settings form(s). The UI splits these fields out into 5 different subforms and reverting one (issuing a DELETE on /api/v2/settings/ldap) would revert all the ldap settings which I think is undesirable.
This fix has the added benefit of cleaning up future activity stream entries for reverting an entire settings category.
I also consulted with @wenottingham and @gamuniz on some fields that were present in the api but missing in the ui and I added those.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
UI
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Add support for Insights as an inventory source
SUMMARY
Use the insights inventory plugin as an inventory source.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
ADDITIONAL INFORMATION
Requires at minimum RedHatInsights/ansible-collections-insights#14
Reviewed-by: Bill Nottingham <None>
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Chris Meyers <None>
Control plane EE
SUMMARY
Per discussion, we are now going to have privileged execution environments for use on the control plane, and these will not be allowed to be edited or deleted.
related ansible/tower#5016
TODO
reinstate the restrictive RBAC for managed_by_tower=True EEs
convert any EEs automatically installed prior into managed_by_tower=False EEs
change the resolver so that ordinary jobs do not get a managed EE, and project updates only get a managed EE
allow sysadmin users to edit the pull field for managed EEs
automatically disassociate EEs that get deleted from the DEFAULT_EXECUTION_ENVIRONMENT setting #10363
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Seth Foster <None>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Resolves issue where FireFox users could not schedule a job template
SUMMARY
This addresses #10152 and #6214
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
ADDITIONAL INFORMATION
Reviewed-by: Michael Abashian <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Adds version context processor back in to fix api browser doc link
SUMMARY
Here's what it looks like on devel now (note the URL in the bottom left):
Here's what it looks like after the change (note the URL in the bottom left):
I dropped this in as it was before the removal of the old UI. I believe the new UI needs access to some of these variables as well to force assets to be refetched after upgrade.
Also of note: I have no idea what I'm doing with django so please help me to become educated if I've done something silly here.
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Stops infinte REST calls on ResourcesAccessAdd workflow
SUMMARY
resolves#10350
There is no functionality changes, I just moved the location of the variables so they are defined outside of the component.
E2E test triggered
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Do not show `not` as choice for Advanced Search on Smart Inventory
Do not show not set type as a choice for Advanced Search on Smart Inventory since
this feature is not implemented on the API side yet.
See: #2817
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Kersom <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Update translation make target for ui_next
SUMMARY
This adds the --clean flag to the extract-strings command (cleans up obsolete strings) and separates out the extract-template command.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
Translation Automation
AWX VERSION
19.2.0
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Christian Adams <rooftopcellist@gmail.com>
Reviewed-by: Alex Corey <Alex.swansboro@gmail.com>
Extend launched by details to handle scheduled JTs and WFJTs
SUMMARY
Issue: #5057
Launched By detail should link to the schedule that launched the job.
ISSUE TYPE
Enhancement Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Reviewed-by: Sarah Akus <sarah.akus@gmail.com>
Add OPTIONS documentation for new job limit feature
Looking at the docs and stuff from #10023
I'm sure this is somewhere else too, but this is the place that users should naturally expect it to be.
Reviewed-by: Chris Meyers <None>
Intentionally drop job event websocket messages in excess of 30 per second (configurable)
SUMMARY
The UI no longer follows the latest job events from websocket messages. Because of that, there's no reason to send messages for all events if the job event rate is high.
I used 30 because this is the number of events that I guesstimate will show in one page in the UI.
Needs the setting added in the UI.
This adds skip_websocket_message to event event_data. We could promote it to a top-level key for job events, if that is preferable aesthetically. Doing this allows us to test this feature without having to connect a websocket client. Ping @mabashian @chrismeyersfsu
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
UI
ADDITIONAL INFORMATION
Scenario walkthrough:
a job is producing 1,000 events per second. User launches it, the screen fills up in, say 1/4 of a second. The scrollbar indicates content beyond the bottom of the screen. Now, for 3/4ths of a second, the scrollbar stays still. After that, it updates the scrollbar to the current line number that the job is on. The scrollbar continues to update the length of the output effectively once per second.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Chris Meyers <None>
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Enter flake8 rules in tox config as actual config entries
I really like to do things like
flake8 awx/main/models
but this is made impractical where running flake8 for the project carries with it an assumption that you will pass a specific set of options.
https://flake8.pycqa.org/en/latest/user/configuration.html
Blessedly, flake8 is flexible in their config options so that it may weather the storm of python project config wars.
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Add a field for hosts automated across to the subscription info
SUMMARY
This is populated by the new table we've added.
Update the subs check to check against this, not imported hosts.
ISSUE TYPE
Feature Pull Request
Bugfix Pull Request
COMPONENT NAME
API
UI
Reviewed-by: Bill Nottingham <None>
Reviewed-by: Amol Gautam <amol_gautam25@yahoo.co.in>
Reviewed-by: Alex Corey <Alex.swansboro@gmail.com>
Reviewed-by: Chris Meyers <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Only attempt to reconnect socket if connection wasn't closed cleanly
SUMMARY
link #8596
After some investigation, I believe that this error is caused by the reconnect logic that we have. When the component that references the ws hook unmounts, we disconnect the socket. There's some logic in the onclose method that attempts to reconnect the socket but if we've disconnected cleanly we shouldn't try to reconnect (because we probably meant to disconnect in the first place).
This should clean up the console errors that we've been seeing about the socket already being in a connected state since we won't have timers running past the lifecycle of the component.
cc @keithjgrant does this sound good to you?
Here's the spec for the disconnect event https://developer.mozilla.org/en-US/docs/Web/API/CloseEvent#properties
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Adds ability to refresh project revision on sync'd rows in the project list
SUMMARY
This PR also adds the revision to the project details view as well as handles updating the revision on the project details view when the project is done syncing.
Since the project revision is not included in the websocket payload, when a project is done syncing the displayed revision may be out of date. As such, we wanted to expose that information to the user and give them the ability to "refresh" and fetch the new revision.
Here's what that flow looks like:
When a particular row finishes syncing the user should see this in place of the revision:
Clicking on that refresh button goes out and fetches the updated project (and with it the potentially updated revision).
We don't do this automatically on the projects list (and force the user to click on the refresh button) is due to issues we've had in the past with the UI triggering API calls based on websocket events.
The flow when a user is on the project details view is a little different because I wasn't as worried about spamming the API with requests.
When a project finishes syncing and the user is viewing the details I do go ahead and automatically refresh the project data (and with it, the revision). Here's what that looks like:
A few other notes:
@tiagodread @akus062381 @one-t I'm almost certain this is going to break some tests because I removed the ClipboardCopyButton in favor of PF's ClipboardCopy component. This component looks and behaves slightly differently from our home grown solution. Additionally, the PF ClipboardCopy button does not expose the ouiaId prop so I had to add a data-cy on that component. You'll likely have to use that identifier to then grab the button inside in order to test out the clipboard copy functionality.
Source Control Revision is a net-new detail in the project details. I think this was simply missed on our initial build of this page.
Here are the identifiers on the various bits:
Note that the identifiers on the project rows have the id of the project appended to them for uniqueness.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
UI
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Michael Abashian <None>
Reviewed-by: Sarah Akus <sarah.akus@gmail.com>
Avoid reload on manual logout
SUMMARY
#10383
Instead of reloading, set the authRedirectTo context variable to "/logout" and handle it as a special case when routing
Reviewed-by: Marliana Lara <marliana.lara@gmail.com>
Remove uses of ansible_virtualenv_path
Run a job with recent AWX, then inspect:
In [2]: Job.objects.order_by('-created').first().custom_virtualenv
Out[2]: '/var/lib/awx/venv/ansible'
We shouldn't do this. The job was ran after the custom_virtualenv field has become entirely informative / non-functional. It is very arguable whether jobs should keep this field. For pre-migration jobs, it is an accurate record of fact, so I'm not going out of my way to argue for removal.
With this change, new jobs will record a blank value for custom_virtualenv, which is more correct, because we don't do this.
I have carefully looked over @rebeccahhh's #10090, and I don't see any way that ansible_virtualenv_path is used. Model use is reported for templates, so that doesn't block removing the field on jobs either.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Rebeccah Hunter <rhunter@redhat.com>
Reviewed-by: Bill Nottingham <None>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Disable cancel button on http 405
SUMMARY
for #9905
When a 405 is received, it means the job is already cancelled. Treat the
request as a success and don't show an error modal.
We disable the button to handle a rare scenario where we receive the 405 long before
the job status is properly updated to "cancelled" over websockets. We want to prevent
more cancel requests when this happens. We're disabling instead of hiding the button
because, assuming the status hasn't changed over sockets, the buttons that usually
replace the cancel button on the toolbar won't be shown yet and we don't want to
needlessly flicker and shift button positions around by rapidly loading and unloading
a different number of buttons onto the bar.
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Sarah Akus <sarah.akus@gmail.com>
Update message for missing EE for Job pages
Update message for missing EE for Job pages
See: #10355
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Kersom <None>
Reviewed-by: Michael Abashian <None>
add a new awx-manage command `custom_venvs`
add an awx-manage command that gets pip freeze data from custom_venv and outputs to command line stdout
SUMMARY
part of #7062 - this command is a glorified pip freeze + some extra stuff, people could navigate to each of their custom virtual environments themselves and run a pip freeze, but this allows them to not, and everyone likes their life to be easier. The extra stuff allows users to see the connections that their existing virtual envs have in awx to things like organizations, jobs, inventory updates, and projects.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.1.0
ADDITIONAL INFORMATION
This is built off of existing code and there is a line that gets custom venv paths from the settings module, that line does not seem to be working. I have written around that but want to make a note of it.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Rebeccah Hunter <rhunter@redhat.com>
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Elijah DeLee <kdelee@redhat.com>
Db partition analytics cmeyers2
Keep old primary key based analytics gathering for unpartitioned
tables.
Use created time on new partitioned tables.
80 million partitioned + 1.5 million unpartitioned Events
Query
awx-manage gather_analytics --dry-run Time
Micro Benchmark Query Time*
Query Only Time**
sequential index scan, multiple ::json casts, 100,000 event batches
102m7.836s
6s
80 minutes
sequential index scan, optimized json cast, 100,000 event batches
48m9.276s
2.2s
30.4 minutes
sequential index scan, optimized json cast, 1,00,000 event batches
39m35.094s
10s
13.3 minutes
sequential table scan, optimized json cast, per-partition batch 600,000 ***
36m42.081s
11.5s
25.5 minutes
*micro benchmarking consists of simply copying a query, running it manually, and observing the runtime.
**micro benchmark time x (80 million / batch size)
**Note that this testing does NOT include the extra modified range query that is needed for correctness. We expect this to be quite fast and is only needed to catch edge case events.
Reviewed-by: Ladislav Smola <lsmola@redhat.com>
Reviewed-by: Elijah DeLee <kdelee@redhat.com>
Make IG credential linkable
SUMMARY
Make the instance group credential detail link to the cred
Reviewed-by: Kersom <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Adds support for typing values into single select lookups
SUMMARY
link #6073
This adds support for typing values into single select lookups. Multi-select lookups (credentials on the JT form, instance groups, etc) will remain unchanged.
An input is now rendered next to the lookup button which will allow users to type a value in.
Here's the button (which opens the modal):
Here's the input (which lets users type a value in):
There's a debounce on the input so that we only make a request to check to see if the name is valid after a second of no typing.
The tricky part of this implementation was handling validation on the lookup fields. If a user types in a string that does not exactly match a value that they can use then we want to show an error indicating that no matching value was found:
but this was tricky because some of these fields are required so the validation functions would need to change a bit.
Since the typed text and the actual underlying value of the lookup are different the validation function would need to have access to both values whenever it runs for this to work. This meant either storing the typed text up at the form level or pushing the validation down into the lookup(s). After talking with @keithjgrant we decided to go down the route of pushing the validation down to the lookups.
You can now pass an optional validation function down to the various lookups that will get combined with the typeahead validation via the useField hook.
Here's the whole thing in action:
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Michael Abashian <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Fixes bug where checkbox list item was selecting things twice
SUMMARY
Resolves#10338
There was a click event on the row and the underlying checkbox. I got rid of the underlying click event so now its only handled at the row level.
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Alex Corey <Alex.swansboro@gmail.com>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Expose the EE resolver to the API
SUMMARY
This change exposes the output of .resolve_execution_environment() for job templates as a new summary field in the API. Note that this is only exposed for the detail views, not the list views. Also, there is the caveat for job templates that may be run under workflows that the workflow EE might override the results of this value.
related #10210
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
AWX VERSION
Reviewed-by: Shane McDonald <me@shanemcd.com>
Reviewed-by: Christian Adams <rooftopcellist@gmail.com>
Reviewed-by: Chris Meyers <None>
Reviewed-by: Jeff Bradberry <None>
Skip sliced jobs from the workflow EE logic
Addressing some unanticipated fallout from #10305
Sliced jobs rely on creating a workflow job, but they do not have a workflow_job_template, so in those cases, that was causing a traceback.
2021-06-03 20:10:52,319 ERROR [a17ebd7f] awx.main.dispatch Worker failed to run task awx.main.tasks.RunJob(*[341], **{}
Traceback (most recent call last):
File "/awx_devel/awx/main/dispatch/worker/task.py", line 90, in perform_work
result = self.run_callable(body)
File "/awx_devel/awx/main/dispatch/worker/task.py", line 65, in run_callable
return _call(*args, **kwargs)
File "/awx_devel/awx/main/tasks.py", line 759, in _wrapped
return f(self, *args, **kwargs)
File "/awx_devel/awx/main/tasks.py", line 1264, in run
self.instance = self.update_model(self.instance.pk, execution_environment=self.instance.resolve_execution_environment())
File "/awx_devel/awx/main/models/mixins.py", line 477, in resolve_execution_environment
if wf_template.execution_environment is not None:
AttributeError: 'NoneType' object has no attribute 'execution_environment'
that is fixed with this, at least for my one simple test case
This left jobs hanging out in "waiting" status, which is really not good.
Reviewed-by: Jeff Bradberry <None>
Add additional controller directory for collections for inventory update
SUMMARY
This may be present in some scenarios with additional collections for inventory usage.
Reviewed-by: Alan Rominger <arominge@redhat.com>
Fix error msg wording and sdb docs
SUMMARY
Fixes wording in container groups error message
Updates ports for sdb debugging docs
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
Reviewed-by: Shane McDonald <me@shanemcd.com>
updated ansible-runner version from 2.0.0a2 -> 2.0.0a4 in requirement…
SUMMARY
version bump (ansible-runner 2.0.0a2 -> 2.0.0a4)
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
API
AWX VERSION
awx: 19.2.0
Reviewed-by: Shane McDonald <me@shanemcd.com>
Fixes bug where user/team role add modal state is not cleared on close
SUMMARY
link #8824
I modeled these changes after the pattern that existed between ResourceAccessList and AddResourceRole. The key to fixing this bug is that the component that renders the wizard needs to be unmounted when the wizard closes so that the state, etc get cleared out before the next time the wizard is opened. In order to achieve that I needed to decouple the add button from the wizard.
The sort of weird part of this pattern (and this exists in the ResourceAccessList as well) is error handling. We pass the error back and set that to state before rendering the modal which isn't quite as clean as having the request made out at the list level and leveraging our hooks for error handling but I decided to just get in and get out and not worry about trying to refactor too much.
Here it is in action:
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Force fully qualified image names
If we try and pull an unqualified image name, jobs hang on a podman
prompt.
I set the permissions as 644 because thats what worked for me because rootless podman needs to be able to read the file, but maybe there is another way to achieve that
Reviewed-by: Christian Adams <rooftopcellist@gmail.com>
allow singular execution_environment for cli
this makes "execution_environments" or "execution_environment" work
which is helpful
Now these both work:
awx execution_environments create --name foobar --image quay.io/ansible/awx-ee:latest
awx execution_environment create --name foobar1 --image quay.io/ansible/awx-ee:latest
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Reviewed-by: Jeff Bradberry <None>
Restore UI license checker
27219d3 removed the license checker for UI dependencies as part of removing the UI. We actually still need it, just pointed at the new UI.
This brings it back, and updates it to do so. It also removes no longer needed license files.
It doesn't pass yet, as I haven't added licenses for the new UI dependencies in this PR. Hi @jakemcdermott.
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Clear list selections on pagination/search/sort
SUMMARY
Updates nearly every list* so that URL param changes (pagination, search, or sort) clear the selection. This prevents the list of selected items in state from including items that may no longer appear on screen — preventing the user from accidentally deleting or otherwise altering an item they may not realize they still have selected.
This also updates the useSelected hook to provide selectAll and clearSelected functions. Any lists that weren't yet already using this hook have been updated to do so.
Addresses #6853 and #7509
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
ADDITIONAL INFORMATION
*Lists that do not include this change are modals where the user is expected to paginate through screens and make several selections along the way (e.g. Multi Credential select modal), and lists that still use PaginatedDataList and are yet to be converted to PaginatedTable
Note: I originally wanted to make the clearSelected prop on PaginatedTable required, so any lists that don't have this fix applied would fail loudly. Unfortunately that wasn't possible, as there were a few lists that should not have this behavior, so I had to leave it as an optional prop.
Reviewed-by: Alex Corey <Alex.swansboro@gmail.com>
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Add Session context and redirects
SUMMARY
Issues:
#8272#9197#8155#9311
When user is logged out due to inactivity, redirect to the Login Page with a warning alert stating that their session expired.
When a user logs in on one tab, redirect from login page to home on the other tabs.
Clear session state when it expires and/or when a user logs out.
Redirect users to direct link destination after authentication
@AlexSCorey is looking at possibly rolling #9915 (SSO login redirect override URL is not working) into this PR
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Alex Corey <Alex.swansboro@gmail.com>
Reviewed-by: Marliana Lara <marliana.lara@gmail.com>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
fix namespaced url params
SUMMARY
Fixes issues when multiple lists are on the page (generally only occurs now when on a list page with an open modal that contains a second list) — Navigating within the modal list currently wipes out URL parameters for the main list, causing the page to reload, which closes the modal. The fix prevents changes to one set of namespaced URL parameters from wiping out URL parameters from another namespace
Refactors query string utils to consolidate a lot of repeated logic from components into a new util, updateQueryString. Use this function to modify query string parameters while maintaining params for multiple namespaces at the same time.
QS utils that are no longer needed have been deleted: replaceParams and encodeNonDefaultQueryString
Addresses #10181 and #9218
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: John Mitchell <None>
Reviewed-by: Alex Corey <Alex.swansboro@gmail.com>
Adds ouiaId to YAML and JSON buttons in Variables Detail
small change to add a better locator to these buttons. if we do not do this, our locators for our visual tests are not good and will eventually break tests
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
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>
when sharing paths use little z
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.
Fixes errors like below:
fatal: [localhost]: FAILED! => {"changed": false, "cmd": "/usr/bin/git ls-remote file:///opt/tmpawx/at_DrunkMail525450112299457413919634186288881628802211907645041298254_test/ -h refs/heads/HEAD", "msg": "fatal: '/opt/tmpawx/at_DrunkMail525450112299457413919634186288881628802211907645041298254_test/' does not appear to be a git repository\\nfatal: Could not read from remote repository.\\n\\nPlease make sure you have the correct access rights\\nand the repository exists.", "rc": 128, "stderr": "fatal: '/opt/tmpawx/at_DrunkMail525450112299457413919634186288881628802211907645041298254_test/' does not appear to be a git repository\\nfatal: Could not read from remote repository.\\n\\nPlease make sure you have the correct access rights\\nand the repository exists.\\n", "stderr_lines": ["fatal: '/opt/tmpawx/at_DrunkMail525450112299457413919634186288881628802211907645041298254_test/' does not appear to be a git repository", "fatal: Could not read from remote repository.", "", "Please make sure you have the correct access rights", "and the repository exists."], "stdout": "", "stdout_lines": []}```
Reviewed-by: Shane McDonald <me@shanemcd.com>
Adds resource name and description to launch and schedule prompt wizards
SUMMARY
link #8249
link #7254
Launch with description:
Launch without description:
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
UI
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Kersom <None>
Reviewed-by: Michael Abashian <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
clarify format of date for host_metric
SUMMARY
@amolgautam25 @AlanCoding let me know what you think -- I was taking a second look at the CLI and took me a bit to remember the format it wants is YYYY-MM-DD
Reviewed-by: Alan Rominger <arominge@redhat.com>
Reviewed-by: Amol Gautam <amol_gautam25@yahoo.co.in>
Fixes bug where users were unable to click on text next to checkboxes/radios in modals
SUMMARY
link #8852
This should impact lists in modals where the user can select one or more of the rows. They should now be able to click on the text/row in order to select. Examples:
ISSUE TYPE
Bugfix Pull Request
COMPONENT NAME
UI
Reviewed-by: Kersom <None>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
Add the ability to disable local authentication
SUMMARY
When an external authentication system is enabled, users would like the ability to disable local authentication for enhanced security.
related #4553
TODO
create a configure-Tower-in-Tower setting, DISABLE_LOCAL_AUTH
expose the setting in the settings UI
be able to query out all local-only users
User.objects.filter(Q(profile__isnull=True) | Q(profile__ldap_dn=''), enterprise_auth__isnull=True, social_auth__isnull=True)
see: awx/main/utils/common.py, get_external_account
write a thin wrapper around the Django model-based auth backend
update the UI tests to include the new setting
be able to trigger a side-effect when this setting changes
revoke all OAuth2 tokens for users that do not have a remote
auth backend associated with them
revoke sessions for local-only users
ultimately I did this by adding a new middleware that checks the value of this new setting and force-logouts any local-only user making a request after it is enabled
settings API endpoint raises a validation error if there are no external users or auth sources configured
The remote user existence validation has been removed, since ultimately we can't know for sure if a sysadmin-level user will still have access to the UI. This is being dealt with by using a confirmation modal, see below.
add a modal asking the user to confirm that they want to turn this setting on
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
UI
AWX VERSION
Reviewed-by: Jeff Bradberry <None>
Reviewed-by: Bianca Henderson <beeankha@gmail.com>
Reviewed-by: Mat Wilson <mawilson@redhat.com>
Reviewed-by: Michael Abashian <None>
Reviewed-by: Chris Meyers <None>
Remove explicit eula acceptance
SUMMARY
Issue: ansible/tower#4948
This PR changes the entitlement process to no require explicit acceptance of the Eula, but still display it for the user to read.
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
API
UI
AWX VERSION
19.1.0
ADDITIONAL INFO
If the eula_accepted field is passed, it will just be ignored.
Reviewed-by: Bill Nottingham <None>
Redirect documentation to awx-operator docs
SUMMARY
CC: @shanemcd
Related: ansible/awx-operator#330
The PR ansible/awx-operator#330 enhances the AWX installation process via the awx-operator.
This PR redirects the installation notes to the awx-operator page so we can simplify and easy maintain the docs in one place.
ISSUE TYPE
Docs Pull Request
Reviewed-by: Shane McDonald <me@shanemcd.com>
The list secrets role rule is no longer not needed for container groups
SUMMARY
The list secrets role rule is no longer not needed for container groups. This was simplified to just get the single matching secret using the read_namespaced_secret() method.
Reviewed-by: Seth Foster <None>
Converts Options lists and other modal lists to tables
SUMMARY
This addresses #10162. Note this issue I found while working on this #10181
Kicked Off E2E
ISSUE TYPE
Feature Pull Request
COMPONENT NAME
UI
AWX VERSION
ADDITIONAL INFORMATION
Reviewed-by: Jake McDermott <yo@jakemcdermott.me>
Reviewed-by: Keith Grant <keithjgrant@gmail.com>
Reviewed-by: Tiago Góes <tiago.goes2009@gmail.com>
