Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[pull] master from apache:master #33

Merged
merged 19 commits into from
Dec 13, 2024
Merged

Conversation

pull[bot]
Copy link

@pull pull bot commented Oct 13, 2024

See Commits and Changes for more details.


Created by pull[bot]

Can you help keep this open source service alive? 💖 Please sponsor : )

Copy link

codecov bot commented Oct 13, 2024

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.50%. Comparing base (86dd867) to head (44705ae).

Additional details and impacted files
@@            Coverage Diff             @@
##           master      #33      +/-   ##
==========================================
- Coverage   81.53%   81.50%   -0.04%     
==========================================
  Files          97       97              
  Lines       12625    12625              
==========================================
- Hits        10294    10290       -4     
- Misses       2050     2052       +2     
- Partials      281      283       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@pull pull bot added the ⤵️ pull label Oct 14, 2024
pbacsko and others added 18 commits October 17, 2024 15:50
Placeholder tracking data is maintained inside the application for
scheduling. If the placeholder is released we update the counters in the
tracking data. We have cases in which we do not do that correctly:
* placeholders are smaller than the real allocation
* placeholder does not have an allocation to replace
* all allocations are removed from an application

Closes: #986

Signed-off-by: Craig Condit <[email protected]>
While releasing placeholders the allocated placeholders are counted
twice in the tracking information. With YUNIKORN-2926 in place this
happens if some but not all placeholders are allocated only.

Mark unallocated placeholders that timeout as released to prevent
scheduling issues.

Closes: #991

Signed-off-by: Craig Condit <[email protected]>
The REST calls to get a statedump and the stacks of the system are
exposed on the wrong endpoint. They belong in /debug and not in /ws/v1.
Both are for troubleshooting only and the content is not stable.

Placing a 301 redirect on the old endpoints for clients to follow.
The proxy build into the web UI does not proxy the debug endpoints.

Closes: #992

Signed-off-by: Craig Condit <[email protected]>
…993)

Remove the Resource object from the REST response for the user and group
trackers. Removed the usage of the DAO objects in the internal unit
tests. The unit tests use new functions which directly expose the same
structures without using the DAO.

Cleanup:
- remove export from internal functions (UserTracker, GroupTracker)
- renamed functions to match their results (Manager)
- update handler to call renamed functions in Manager
- reimplemented test functions to not use DAO objects (Manager,
  UserTracker, GroupTracker, QueueTracker)
- assert function changes to use new functions removing DAO usage
  (scheduler/utilites_test, scheduler/objects/utilites_test)

Closes: #993

Signed-off-by: Craig Condit <[email protected]>
Only respond with the allowed methods for the request, not with a
general all allowed set. OPTIONS is supported via the generic config.
Add a test to make sure a change in router does not break that.

Remove the Access-Control-Allow-Credentials as recommended in the RFC.
We also do not use cookies or authentication so not relevant to set.

Closes: #994

Signed-off-by: Craig Condit <[email protected]>
…rs (#996)

YUNIKORN-2700 introduced a bug where allocations of previously-reserved
tasks were not handled correctly in the case where we schedule on a
different node than the reservation. Ensure that we unreserve and
allocate using the proper node in both cases.

Also introduce additional logging of allocations on nodes to make
finding issues like this easier in the future.

Closes: #996
Remove the TrackedResource object from the REST response for the
application. Moved the tracked resources into a separate object inside
the app DAO, as we have done for other complex pieces.
Cleanup exposure of the summary object to be testing only.
Add DAO retrieval of TrackedResource type for the web handler.

Closes: #999

Signed-off-by: Craig Condit <[email protected]>
Add preemption and priority details to the queue REST information.
Currently the configuration is exposed via the properties. The
properties contain the raw configuration values which need to be
interpreted.

Invalid configuration values are ignored and not set on the queue. This
could cause an incorrect view of what is or should be active on the
queue if based on the properties. Using the active values from the
object to show the real state.

Expose Partition level preemption flag as part of this change to
complement the queue details.

Closes: #995

Signed-off-by: Craig Condit <[email protected]>
@chenyulin0719 chenyulin0719 merged commit f7d0e10 into chenyulin0719:master Dec 13, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

8 participants