Commit Graph

3925 Commits

Author SHA1 Message Date
Kubernetes Prow Robot
4a1558c545 Merge pull request #133967 from pohly/dra-allocator-selection
DRA: allocator selection
2025-09-30 08:24:18 -07:00
Patrick Ohly
60eeaa6ebd DRA scheduler: add unit test for allocator selection
This prevents the mistake from 1.34 where the default-on
DRAResourceClaimDeviceStatus feature caused the use of the experimental
allocator implementation. The test fails without a fix for that.
2025-09-30 16:53:38 +02:00
Patrick Ohly
7f57730ba4 DRA scheduler: fix selection of "incubating" allocator implementation
In 1.34, the default feature gate selection picked the "experimental" allocator
implementation when it should have used the "incubating" allocator. No harm
came from that because the experimental allocator has all the necessary if
checks to disable the extra code and no bugs were introduced when implementing
it, but it means that our safety net wasn't there when we expected it to be.

The reason is that the "DRAResourceClaimDeviceStatus" feature gate is on by
default and was only listed as supported by the experimental implementation.
This could be fixed by listing it as supported also by the other
implementation, but that would be a bit odd because there is nothing to support
for it (the reason why this was missed in 1.34!). Instead, the allocator
features are now only indirectly related to feature gates, with a single
boolean controlling the implementation of binding conditions.
2025-09-30 16:53:38 +02:00
Patrick Ohly
b5bcac998d DRA scheduler: clean up feature gate handling
Copying from feature.Features to new fields in the plugin got a bit silly with
the long list of features that we have now. Embedding feature.Features is
simpler.

Two fields in feature.Features weren't named according to the feature gate, now
they are named consistently and the fields are sorted.
2025-09-30 16:53:38 +02:00
Maciej Skoczeń
d2e6be440c Revert "Merge pull request #133213 from sanposhiho/second-trial-conor"
This reverts commit a2bf45b081, reversing
changes made to 2b2ea27250.
2025-09-24 11:05:16 +00:00
Maciej Skoczeń
4275b8b759 Don't limit the number of goroutines dispatched by the API Dispatcher 2025-09-19 09:36:44 +00:00
hojinchoi
7028ba09db fix: duplicated 'the' in comment 2025-09-18 18:11:44 +09:00
Lalit Chauhan
af6b3ed1f5 Add support for k8s-label-key 2025-09-12 16:52:04 +00:00
Kubernetes Prow Robot
d602326b87 Merge pull request #133363 from yliaog/implicit
Allow implicit extended resource name to be used no matter explicit extendedResourceName field is set or not in device class
2025-09-11 13:40:07 -07:00
yliao
74cf1db218 sort the device requests in the extended resource claim spec.
removed the sortClaim in the unit test.
2025-09-11 16:55:58 +00:00
yliao
79f8d1b1c5 fixed bug such that implicit extended resource name can always be used,
no matter the explicit extendedResourceName field in device class is set or not.
2025-09-10 14:10:40 +00:00
Maciej Skoczeń
3dfcda9afd Fix minor inconsistencies in scheduler 2025-09-10 11:40:10 +00:00
Kubernetes Prow Robot
597a684bb0 Merge pull request #133172 from ania-borowiec/move_handle_and_plugin
Move interfaces: Handle and Plugin and related types from kubernetes/kubernetes to staging repo kube-scheduler
2025-09-08 06:05:31 -07:00
Maciej Skoczeń
4babdf8026 Fix race in movePodsToActiveOrBackoffQueue 2025-09-02 11:57:18 +00:00
Ania Borowiec
fadb40199f Move interfaces: Handle and Plugin and related types from kubernetes/kubernetes to staging repo kube-scheduler 2025-09-02 09:42:53 +00:00
Kubernetes Prow Robot
b94b6ece10 Merge pull request #133707 from ania-borowiec/fitsports
Remove use of pkg/scheduler/framework.NodeInfo in node_ports.go
2025-08-31 19:25:11 -07:00
Kubernetes Prow Robot
5fb3296920 Merge pull request #132451 from macsko/fix_race_in_scheduler_integration_tests
Fix race in scheduler integration tests
2025-08-31 05:03:09 -07:00
Kubernetes Prow Robot
871857b0d0 Merge pull request #133608 from yliaog/flake
added resourceClaimModified to bindClaim to decide whether to update assume cache
2025-08-29 15:23:08 -07:00
yliao
bf13cd1b81 added resourceClaimModified to bindClaim to decide whether to update assume cache 2025-08-29 16:12:55 +00:00
Kubernetes Prow Robot
4d97cda614 Merge pull request #133710 from ania-borowiec/affinityterms
Move GetAffinityTerms functions to staging repo
2025-08-27 21:39:32 -07:00
Kubernetes Prow Robot
4d926683ca Merge pull request #133518 from macsko/dont_run_preenqueue_when_pod_is_activated_from_backoffq
Don't run PreEnqueue when pod is activated from backoffQ
2025-08-27 18:30:38 -07:00
Kubernetes Prow Robot
c725a54773 Merge pull request #133360 from sanposhiho/flake-shced-on
fix: parsing the patch directly
2025-08-27 17:18:11 -07:00
Kubernetes Prow Robot
36a1587dbf Merge pull request #133080 from togettoyou/cleanup-scheduler_perf_TODO
scheduler_perf: remove outdated TODO comments
2025-08-27 16:05:27 -07:00
Ania Borowiec
b012e16b47 Remove use of pkg/scheduler/framework.NodeInfo in node_ports.go 2025-08-27 13:30:45 +00:00
Ania Borowiec
3c00c3cb29 Move GetAffinityTerms functions from pkg/scheduler/framework to staging repo 2025-08-26 13:39:49 +00:00
Maciej Skoczeń
46e10103ff Take activeQ lock for part of the Update method 2025-08-25 12:30:43 +00:00
Kensei Nakada
3bf684b4ed fix: parsing the patch directly 2025-08-24 23:41:14 +09:00
Maciej Skoczeń
8b0b0df431 Don't run PreEnqueue when pod is activated from backoffQ 2025-08-22 12:40:41 +00:00
Abu Kashem
747a295cac fix flake in dra test 'TestPlugin'
TestPlugin/multi-claims-binding-conditions-all-success/PreEnqueue
flakes due to the assumed cache not been synced with the initial
store. The test waits until the registered handler used by the
assumed cache has synced before proceeding with the test
2025-08-18 15:54:03 -04:00
Junhao Zou
a405844e66 scheduler_perf(fix): correct typos, clarify comments, and minor tweaks 2025-08-15 16:34:38 +08:00
Abu Kashem
c8ab780edb dra plugin: assume claim after api call in bindClaim 2025-08-13 16:35:35 -04:00
yliao
2a026f6d65 1/ added retries to AssumeClaimAfterAPICall for the object which is not present in the cache (dynamicresources.go)
2/ modified the assume cache verification to not error out as long as
the expected claim is in the cache, no matter its latest and api object
are different or not. (dynamicresources_test.go).
3/ fixed nil panic as seen from https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/133321/pull-kubernetes-integration/1952472629470302208
2025-08-06 07:08:58 +00:00
Maciej Skoczeń
aa59f930b3 Add lock to TestAsyncPreemption to prevent races 2025-08-05 09:43:12 +00:00
Maciej Skoczeń
c5ef720837 Fix race in scheduler integration tests 2025-08-05 09:42:52 +00:00
Kubernetes Prow Robot
b37978f226 Merge pull request #133334 from macsko/fix_potential_race_in_patchpodstatus_api_call_implementation
Fix potential race in PodStatusPatchCall implementation
2025-08-01 02:47:45 -07:00
Kubernetes Prow Robot
ea81dd6d01 Merge pull request #133309 from macsko/fix_race_with_closing_api_dispatcher
Fix potential race in closing API dispatcher
2025-08-01 02:47:38 -07:00
Maciej Skoczeń
9eda4789c0 Fix potential race in PodStatusPatchCall implementation 2025-07-31 09:27:40 +00:00
Maciej Skoczeń
dbfeb9c351 Fix potential race in closing API dispatcher 2025-07-30 11:57:26 +00:00
Kensei Nakada
46b858aa13 fix: return false to apply the patch 2025-07-30 19:12:57 +09:00
yliao
0a12f00e9d fix nil panic in hasBindingConditions, it cannot assume claim has allocations 2025-07-30 14:44:41 +09:00
Sunyanan Choochotkaew
7f052afaef KEP 5075: implement scheduler
Signed-off-by: Sunyanan Choochotkaew <sunyanan.choochotkaew1@ibm.com>
2025-07-30 09:52:49 +09:00
Sunyanan Choochotkaew
5ad969588d KEP-5075: API updates
Signed-off-by: Sunyanan Choochotkaew <sunyanan.choochotkaew1@ibm.com>
2025-07-30 09:26:40 +09:00
Kubernetes Prow Robot
1b273b385e Merge pull request #130653 from yliaog/master
kubelet and scheduler for extended resource backed by DRA
2025-07-29 13:04:27 -07:00
yliao
34a64db2c7 extended resource backed by DRA: implementation 2025-07-29 18:55:21 +00:00
Kubernetes Prow Robot
74f7a44966 Merge pull request #133276 from macsko/stop_clearing_nnn_in_all_cases
KEP-5278 Stop clearing NominatedNodeName in all cases
2025-07-29 11:24:40 -07:00
Kubernetes Prow Robot
e2ab840708 Merge pull request #130160 from KobayashiD27/dra-device-binding-conditions
Implement DRA Device Binding Conditions (KEP-5007)
2025-07-29 07:34:26 -07:00
Maciej Skoczeń
aea0a3cca2 Run all relevant test cases with the feature gate enabled and disabled 2025-07-29 12:21:03 +00:00
utam0k
856e7d2383 scheduler: Stop clearing NominatedNodeName on all cases
Signed-off-by: utam0k <k0ma@utam0k.jp>
2025-07-29 12:21:03 +00:00
Kobayashi,Daisuke
e8c3af1f5c KEP-5007 DRA Device Binding Conditions: Implement scheduler logic 2025-07-29 11:34:30 +00:00
Kensei Nakada
ac9fad6030 feat: trigger PreFilterPreBind in the binding cycle 2025-07-29 19:01:02 +09:00