github/kubernetes
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
102,527 Commits 1 Branch 31 Tags
9d30eb88dfd1cabc87e08793bcad07e6faec46c2
Commit Graph

42647 Commits

Author SHA1 Message Date
Elana Hashman
b5f24c334e Bump DynamicKubeConfig metric deprecation to 1.23 2021-07-28 09:29:57 -07:00
Kubernetes Prow Robot
47e1df8f4e Merge pull request #103743 from kolyshkin/runc-1.0.1 
vendor: bump runc to v1.0.1
 2021-07-23 15:16:33 -07:00
Jordan Liggitt
236e72cf8a Make CSR cleaner tolerate objects with invalid status.certificate 2021-07-21 10:35:17 -04:00
Kubernetes Prow Robot
9f47110aa2 Merge pull request #103785 from smarterclayton/preserve_reason 
Ensure that Reason and Message are preserved on pod status
 2021-07-20 15:21:26 -07:00
Kubernetes Prow Robot
7303b2ce4e Merge pull request #103747 from mattcary/noapi 
Revert #99378
 2021-07-20 07:19:31 -07:00
Kubernetes Prow Robot
6aa160f3ba Merge pull request #103181 from 249043822/bugfix-volumemanager 
Add sync reconstructed volume from desired state of world for volumemanager
 2021-07-19 15:04:52 -07:00
Clayton Coleman
d7ee024cc5 kubelet: Make condition processing in one spot 
The list of status conditions should be calculated all together,
this made review more complex. Readability only.
 2021-07-19 17:56:22 -04:00
Clayton Coleman
c2a6d07b8f kubelet: Avoid allocating multiple times during status 
Noticed while reviewing this code path. We can assume the
temporary slice should be about the same size as it was previously.
 2021-07-19 17:55:18 -04:00
Clayton Coleman
9efd40d72a kubelet: Preserve reason/message when phase changes 
The Kubelet always clears reason and message in generateAPIPodStatus
even when the phase is unchanged. It is reasonable that we preserve
the previous values when the phase does not change, and clear it
when the phase does change.

When a pod is evicted, this ensurse that the eviction message and
reason are propagated even in the face of subsequent updates. It also
preserves the message and reason if components beyond the Kubelet
choose to set that value.

To preserve the value we need to know the old phase, which requires
a change to convertStatusToAPIStatus so that both methods have
access to it.
 2021-07-19 17:54:55 -04:00
Kir Kolyshkin
e5b434e990 kubelet/cm: don't set Devices 
Since runc 1.0.0 it is now sufficient to have SkipDevices: true.

Signed-off-by: Kir Kolyshkin <kolyshkin@gmail.com>
 2021-07-16 12:45:35 -07:00
Matthew Cary
aeb82243fc Revert "tests for statefulset PersistentVolumeClaimDeletePolicy api change" 
This reverts commit b259686b36.

Change-Id: I25cc8ae8a9aa77d0bc483147d3f8fb65616a2b2b
 2021-07-16 11:51:35 -07:00
Kubernetes Prow Robot
76b0906136 Merge pull request #101047 from jsturtevant/issue-100966-dualstack-windows 
[sig-windows] update winkernel to only use dualstack if the node and config supports it
 2021-07-15 20:30:09 -07:00
Matthew Cary
10a3cc815b Revert "statefulset PersistentVolumeClaimDeletePolicy api change" 
This reverts commit e4e735cd1c.

Change-Id: I3e17aa585fd9ae4f9e74b97326ae6b9fadd2a94c
 2021-07-15 17:14:19 -07:00
Matthew Cary
fb5b966a88 Revert "Add StatefulSetAutoDeletePVC feature gate" 
This reverts commit 0feba0c4ee.
 2021-07-15 17:03:53 -07:00
Kubernetes Prow Robot
02355e22b3 Merge pull request #103650 from dims/get-golangci-lint-back-on-its-feet 
Get golangci lint back on its feet
 2021-07-15 02:06:48 -07:00
Kubernetes Prow Robot
68922e3deb Merge pull request #103644 from alculquicondor/tracking-deleted-pods 
Revert counting deleted pods as failures for Job
 2021-07-14 09:02:28 -07:00
Aldo Culquicondor
5e1b5ec398 Revert counting deleted pods as failures for Job 
When JobTrackingWithFinalizers is disabled. To preserve existing behavior.

Change-Id: Id1752f96feed322911712fe9e918e91e42eca809
 2021-07-14 10:03:20 -04:00
Davanum Srinivas
75748c185e enable verify-golangci-lint.sh 
Signed-off-by: Davanum Srinivas <davanum@gmail.com>
 2021-07-14 08:53:33 -04:00
Davanum Srinivas
07332ad398 fix ineffassign and varcheck 
Signed-off-by: Davanum Srinivas <davanum@gmail.com>
 2021-07-14 08:41:22 -04:00
Davanum Srinivas
26cc8e40a8 fix deadcode issues 
Signed-off-by: Davanum Srinivas <davanum@gmail.com>
 2021-07-14 08:41:21 -04:00
Kubernetes Prow Robot
2da4d48e6d Merge pull request #100567 from jingxu97/mar/mark 
Mark volume mount as uncertain in case of volume expansion fails
 2021-07-13 22:20:26 -07:00
Kubernetes Prow Robot
d6f2473d08 Merge pull request #103668 from smarterclayton/panic_in_pod_worker 
kubelet: Prevent runtime-only pods from going into terminated phase
 2021-07-13 17:42:26 -07:00
Clayton Coleman
de9cdab5ae kubelet: Prevent runtime-only pods from going into terminated phase 
If a pod is already in terminated and the housekeeping loop sees an
out of date cache entry for a running container, the pod worker
should ignore that running pod termination request. Once the worker
completes, a subsequent housekeeping invocation will then invoke
terminating because the worker is no longer processing any pod
with that UID.

This does leave the possibility of syncTerminatedPod being blocked
if a container in the pod is started after killPod successfully
completes but before syncTerminatedPod can exit successfully,
perhaps because the terminated flow (detach volumes) is blocked on
that running container. A future change will address that issue.
 2021-07-13 15:41:49 -04:00
Elana Hashman
aaa7de0ac6 Update API description for probe.terminationGracePeriodSeconds 2021-07-13 14:28:06 -05:00
rarashid
bf2ae14501 Move feature flag to beta (but leave as false) and remove the feature flag from Kubelet 2021-07-13 14:25:44 -05:00
Kubernetes Prow Robot
e375563732 Merge pull request #103245 from wzshiming/fix/prober-termination 
Add validation for Prober TerminationGracePeriodSeconds
 2021-07-13 09:30:30 -07:00
KeZhang
65618bfd69 Add sync reconstructed volume from desired state of world for volumemanager 2021-07-13 12:51:37 +08:00
Kubernetes Prow Robot
04ef2b115d Merge pull request #90216 from DataDog/nayef/fix-container-statuses-race 
Avoid overwriting podStatus ContainerStatuses in convertToAPIContainerStatuses
 2021-07-12 17:02:29 -07:00
Kubernetes Prow Robot
584658a17d Merge pull request #103601 from dashpole/webhook_tracing 
Add tracing to admission webhook requests
 2021-07-12 10:22:47 -07:00
Kubernetes Prow Robot
e799d7b191 Merge pull request #99023 from verb/1.21-securitycontext 
Allow setting securityContext in ephemeral containers
 2021-07-09 20:50:46 -07:00
Kubernetes Prow Robot
de0670ca0d Merge pull request #103553 from ehashman/rename-node-ff 
Rename NodeSwapEnabled flag to NodeSwap
 2021-07-09 16:49:30 -07:00
Kubernetes Prow Robot
60756214d8 Merge pull request #103533 from jiahuif/feature/leader-migration/to-beta 
KEP-2436 Leader Migration: to beta
 2021-07-09 12:15:00 -07:00
Elana Hashman
642eff0c69 Rename NodeSwapEnabled flag to NodeSwap 2021-07-09 11:39:52 -07:00
Kubernetes Prow Robot
2423813207 Merge pull request #103573 from chendave/fix_index 
Fix index out of range if multiple default plugins are overridden
 2021-07-09 08:43:23 -07:00
Kubernetes Prow Robot
a6c2cd7d18 Merge pull request #103291 from wzshiming/fix/nodeshutdown-restart 
Fix Data Race in nodeshutdown restart
 2021-07-09 08:43:14 -07:00
Kubernetes Prow Robot
03fa68099e Merge pull request #98238 from alculquicondor/job-completion 
Track Job completion through Pod finalizers and status
 2021-07-09 08:42:54 -07:00
David Ashpole
9dd59017c4 add tracing to webhook requests 2021-07-09 06:30:05 -07:00
Kubernetes Prow Robot
29652248eb Merge pull request #103596 from andrewsykim/endpointslice-terminating 
Promote EndpointSliceTerminatingCondition to Beta
 2021-07-09 06:01:42 -07:00
Kubernetes Prow Robot
8daced4d3f Merge pull request #103508 from boenn/UseDiff 
Use cmp.Diff() replace reflect and diagnosis
 2021-07-09 06:01:13 -07:00
Dave Chen
1727cea64c Fix index out of range if multiple default plugins are overridden 
Signed-off-by: Dave Chen <dave.chen@arm.com>
 2021-07-09 19:56:14 +08:00
Indeed
995278c9fb add ControllerManagerLeaderMigration as beta 
This feature was in staging/src/k8s.io/controller-manager/pkg/features/kube_features.go but missing here.
 2021-07-08 21:20:13 -07:00
Kubernetes Prow Robot
617064d732 Merge pull request #101432 from swatisehgal/smtaware 
node: cpumanager: add options to reject non SMT-aligned workload
 2021-07-08 21:04:53 -07:00
Shiming Zhang
513bd93f76 update test for feature gate 2021-07-09 10:37:38 +08:00
Kubernetes Prow Robot
83baa708df Merge pull request #103429 from saschagrunert/metrics-test-fix 
Fix resource metrics e2e test
 2021-07-08 17:58:53 -07:00
Kubernetes Prow Robot
dab6f6a43d Merge pull request #102344 from smarterclayton/keep_pod_worker 
Prevent Kubelet from incorrectly interpreting "not yet started" pods as "ready to terminate pods" by unifying responsibility for pod lifecycle into pod worker
 2021-07-08 16:48:53 -07:00
Jing Xu
0fa01c371c Mark volume mount as uncertain in case of volume expansion fails 
should mark volume mount in actual state even if volume expansion fails so that
reconciler can tear down the volume when needed. To avoid pods start
using it, mark volume as uncertain instead of mounted.

Will add unit test after the logic is reviewed.

Change-Id: I5aebfa11ec93235a87af8f17bea7f7b1570b603d
 2021-07-08 16:00:34 -07:00
Kubernetes Prow Robot
57716897eb Merge pull request #103434 from perithompson/windows-etchostcreate-skip 
Explicitly skip host file mounting for Windows when HostProcess pod
 2021-07-08 15:36:53 -07:00
Andrew Sy Kim
826a5219da promote EndpointSliceTerminatingCondition to Beta 
Signed-off-by: Andrew Sy Kim <kim.andrewsy@gmail.com>
 2021-07-08 17:34:10 -04:00
Francesco Romani
23abdab2b7 smtalign: propagate policy options to policies 
Consume in the static policy the cpu manager policy options from
the cpumanager instance.
Validate in the none policy if any option is given, and fail if so -
this is almost surely a configuration mistake.

Add new cpumanager.Options type to hold the options and translate from
user arguments to flags.

Co-authored-by: Swati Sehgal <swsehgal@redhat.com>
Signed-off-by: Francesco Romani <fromani@redhat.com>
 2021-07-08 23:15:37 +02:00
Francesco Romani
6dcec345df smtalign: cm: factor out admission response 
Introduce a new `admission` subpackage to factor out the responsability
to create `PodAdmitResult` objects. This enables resource manager
to report specific errors in Allocate() and to bubble up them
in the relevant fields of the `PodAdmitResult`.

To demonstrate the approach we refactor TopologyAffinityError as a
proper error.

Co-authored-by: Kevin Klues <kklues@nvidia.com>
Co-authored-by: Swati Sehgal <swsehgal@redhat.com>
Signed-off-by: Francesco Romani <fromani@redhat.com>
 2021-07-08 23:15:37 +02:00