Commit Graph

126494 Commits

Author SHA1 Message Date
Kubernetes Prow Robot
ce6396175b
Merge pull request #127318 from aroradaman/conntrack-reconciler
proxy/conntrack: reconciler
2024-10-31 10:21:33 +00:00
Kubernetes Prow Robot
f94f87795f
Merge pull request #126935 from aojea/proxy_conntrack_service_topology
e2e conntrack test for UDP Service with internalTrafficPolicy local
2024-10-31 10:21:26 +00:00
Maciej Szulik
174288d751
Add unit tests verifying the update touches old, unhealthy pods first, and only after new pods.
Signed-off-by: Maciej Szulik <soltysh@gmail.com>
2024-10-31 11:13:01 +01:00
Patrick Ohly
d1f0d24ef9 DRA API: check "AdminAccess in use" only once
This is simpler and an opportunity to explain the concept.
2024-10-31 09:42:18 +01:00
Kensei Nakada
bd8e9dd021 fix: flake TestSchedulerScheduleOne 2024-10-31 17:13:50 +09:00
Jan Safranek
9e29f95618 Refactor controller-manager volume plugins
Most of the volume plugins were removed from k/k. Refactor how KCM
controllers initialize the few leftovers.
2024-10-31 09:08:30 +01:00
Jan Safranek
cba5a93468 Remove portworx from attachable volume plugins
The volume plugin does not implement AttachableVolumePlugin interface.
2024-10-31 09:08:21 +01:00
Jan Safranek
0ecbdf3622 Remove fc from expandable plugins
FibreChannel volume plugin does not implement ExpandableVolumePlugin.
2024-10-31 09:08:21 +01:00
Jan Safranek
1fa8877c33 Add unit tests for KCM volume plugin probers 2024-10-31 09:08:19 +01:00
Kubernetes Prow Robot
453efd7a4b
Merge pull request #121604 from pacoxu/image-pull-e2e
[node-e2e] add test cases for serialize and parallel image pulling
2024-10-31 08:01:26 +00:00
Paco Xu
82df7a7d82 use cri proxy injector for parallel pulling image tests 2024-10-31 14:50:50 +08:00
yunwang0911
f428881ec0 Update pkg/kubelet/status/state/state_checkpoint_test.go 2024-10-31 14:02:35 +08:00
Kubernetes Prow Robot
7c56aa5a58
Merge pull request #128353 from sanposhiho/patch-13
fix: register ResourceSlice to allResources
2024-10-31 04:41:25 +00:00
Kubernetes Prow Robot
5d353417cd
Merge pull request #128346 from dims/update-to-latest-advisor-for-1.32
Update to latest cadvisor - `v0.51.0`
2024-10-30 23:45:26 +00:00
Kubernetes Prow Robot
c0e0785fe4
Merge pull request #128427 from dom4ha/scheduler-perf
Fix Unschedulable test by using high priority churn pods to get processed right after they were injected
2024-10-30 22:23:25 +00:00
Joe Betz
c59fba7f26
Promote CRD field selector e2e test to conformance (#128109)
* Promote CRD field selector e2e test to conformance

* Fix release number for conformance test

* re-run update conformance
2024-10-30 21:19:25 +00:00
Richa Banker
6eeac6d32f Add tests 2024-10-30 14:01:32 -07:00
Richa Banker
8fd128085c do not install handler for /metrics/slis using sync.Once 2024-10-30 14:01:31 -07:00
Kubernetes Prow Robot
dc1d7f41ef
Merge pull request #128456 from benluddy/nondeterministic-response-encoding
KEP-4222: Allow nondeterministic object encoding in HTTP response bodies.
2024-10-30 20:13:27 +00:00
Kevin Conner
9538747d4d Fix elapsed time tracking for validating admission policies
Signed-off-by: Kevin Conner <kev.conner@gmail.com>
2024-10-30 12:38:39 -07:00
Davanum Srinivas
152d342a8d
Update to latest cadvisor
Signed-off-by: Davanum Srinivas <davanum@gmail.com>
2024-10-30 15:25:21 -04:00
Ben Luddy
dee76a460e
Allow nondeterministic object encoding in HTTP response bodies. 2024-10-30 15:10:16 -04:00
Kubernetes Prow Robot
16f9fdc705
Merge pull request #128273 from benluddy/cbor-apply
KEP-4222: Support CBOR encoding for apply requests.
2024-10-30 17:25:25 +00:00
Sergey Kanzhelev
1297d0cdd1 converge DRA and Device Plugin plugins registration 2024-10-30 16:58:13 +00:00
Ben Luddy
41f55d7117
Regenerate clients to support application/apply-patch+cbor. 2024-10-30 12:21:15 -04:00
Ben Luddy
37ed906a33
Support application/apply-patch+cbor in patch requests. 2024-10-30 12:21:15 -04:00
Kubernetes Prow Robot
6435489064
Merge pull request #128275 from pohly/dra-resourceslice-controller-multiple-slices
DRA resourceslice controller: support publishing multiple slices
2024-10-30 16:01:26 +00:00
Joe Betz
af811be190 Don't pick versions that have a replacement as storage version 2024-10-30 11:47:43 -04:00
Oksana Baranova
49b88f1d8a kubelet: migrate clustertrustbundle, token to contextual logging
Signed-off-by: Oksana Baranova <oksana.baranova@intel.com>
2024-10-30 17:31:11 +02:00
Jordan Liggitt
43fa7c638b
Revert "Merge pull request #127669 from olyazavr/fix-probe-race"
This reverts commit 3d00d6e421, reversing
changes made to a7fcc89ac0.
2024-10-30 11:16:47 -04:00
Patrick Ohly
1088f4fb44 DRA resourceslice controller: do DeepCopy for driver resources
The reason for the previous behavior was unnecessary performance overhead that
occurs when the caller already provided a "fresh" copy and doesn't touch it
afterwards.

But this is something that DRA driver developers can easily get wrong, so it's
better to be safe than sorry.
2024-10-30 15:54:32 +01:00
Patrick Ohly
67f0428769 DRA resourceslice controller: delay sync
When deleting a bunch of slices, the delete events queue the pool while it is
being synced. It then got synced again immediately, while the deleted slices
were still being removed from the informer cache. The obsolete slice in the
cache caused the controller to delete it again, which fails with a "not
found". That error is ignored, but this still caused extra API calls.

Now syncing gets delayed with a configuration duration (default: 30 seconds) so
the informer cache is more likely to be up-to-date when the pool gets synced
again.
2024-10-30 15:54:32 +01:00
Patrick Ohly
99cf2d8a2e DRA resource slice controller: add E2E test
This test covers creating and deleting 100 large ResourceSlices. It is strict
about using the minimum number of calls.

The test also verifies that creating large slices works.
2024-10-30 15:54:32 +01:00
Patrick Ohly
7473e643fa DRA resource slice controller: use MutationCache to avoid race
This avoids the problem of creating an additional slice when the one from the
previous sync is not in the informer cache yet. It also avoids false
attempts to delete slices which were updated in the previous sync. Such
attempts would fail the ResourceVersion precondition check, but would
still cause work for the apiserver.
2024-10-30 15:54:32 +01:00
Patrick Ohly
e88d5c37e6 DRA resource claim controller: add statistics
This is primarily for testing. Proper metrics might be useful, but can still be
added later.
2024-10-30 15:54:32 +01:00
Patrick Ohly
d94752ebc8 DRA resourceslice controller: use preconditions for Delete
It's better to verify UID and ResourceVersion of the ResourceSlice that we want
to delete. If anything changed, the decision to remove it might not apply
anymore and we need to check again.
2024-10-30 15:54:32 +01:00
Patrick Ohly
a6d180c7d3 DRA: validate set of devices in a pool before using the pool
The ResourceSlice controller (theoretically) might end up creating too many
slices if it syncs again before its informer cache was updated. This could
cause the scheduler to allocate a device from a duplicated slice. They should
be identical, but its still better to fail and wait until the controller
removes the redundant slice.
2024-10-30 15:54:32 +01:00
Patrick Ohly
26650371cc DRA resourceslice controller: support publishing multiple slices
The driver determines what each slice is meant to look like. The controller
then ensures that only those slices exist. It reuses existing slices where the
set of devices, as identified by their names, is the same as in some desired
slice. Such slices get updated to match the desired state.

In other words, attributes and the order of devices can be changed by updating
an existing slice, but adding or removing a device is done by deleting and
re-creating slices.

Co-authored-by: googs1025 <googs1025@gmail.com>

The test update is partly based on
https://github.com/kubernetes/kubernetes/pull/127645.
2024-10-30 15:54:32 +01:00
Antoni Zawodny
4afa554f65 Add --concurrent-daemonset-syncs flag to kube-controller-manager 2024-10-30 15:03:26 +01:00
dom4ha
ff584a76e0 Fix Unschedulable test by scheduling high priority churn pods to get processed right after they were injected (before the queued test pods) 2024-10-30 13:04:38 +00:00
Itamar Holder
f21473b924 Set pod-level CPUPeriod only if CPUQuota is changed
Signed-off-by: Itamar Holder <iholder@redhat.com>
2024-10-30 14:21:35 +02:00
Itamar Holder
c792c30b6a Refactor: remove no longer needed resourceName parameter
Signed-off-by: Itamar Holder <iholder@redhat.com>
2024-10-30 13:58:38 +02:00
Itamar Holder
7207ce20f0 Refactor: remove functions that are no longer used
Signed-off-by: Itamar Holder <iholder@redhat.com>
2024-10-30 13:58:38 +02:00
Itamar Holder
510ff67528 Use libcontainer's cgroup manager to update resources through systemd
libcontainer's cgroup manager is version agnostic, and is agnostic
to whether systemd is used. This way if systemd is used, the cgroup
manager would be able to update resources properly so that if
the daemon would be restarted the changes would not be reverted.

Signed-off-by: Itamar Holder <iholder@redhat.com>
2024-10-30 13:58:38 +02:00
Itamar Holder
2a5a6c7fb8 Refactor: add import alias to libcontainer cgroup manager
Signed-off-by: Itamar Holder <iholder@redhat.com>
2024-10-30 13:58:38 +02:00
Kubernetes Prow Robot
d001d5684e
Merge pull request #128417 from tenzen-y/self-nominate-job-controller-reviewer
Self nominate tenzen-y as a reviewer for the Job controller
2024-10-30 11:21:39 +00:00
Kubernetes Prow Robot
a18b50e7e4
Merge pull request #128373 from mimowo/job-cover-negative-codes
Job Pod Failure policy - cover testing of negative exit codes
2024-10-30 11:21:31 +00:00
Kubernetes Prow Robot
7529696b59
Merge pull request #128334 from mimowo/job-windows-e2e-test
Job Pod Failure policy refactor e2e test using exit codes
2024-10-30 11:21:25 +00:00
yunwang0911
05493c0924
Update pkg/kubelet/status/state/state_checkpoint_test.go
Co-authored-by: Tim Allclair <timallclair@gmail.com>
2024-10-30 18:11:10 +08:00
yunwang0911
e4c8eefeb2
Update pkg/kubelet/status/state/state_checkpoint_test.go
Co-authored-by: Tim Allclair <timallclair@gmail.com>
2024-10-30 18:08:53 +08:00