profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/endocrimes/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.
Danielle endocrimes Staff Engineer @VMware Berlin, Germany dani.builds.terrible.systems kubernetes @vmware, previously nomad @hashicorp, build platform @circleci, open sourceress @cocoapods

CocoaPods/CocoaPods 13381

The Cocoa Dependency Manager.

CocoaPods/CocoaPods-app 1217

A full-featured and standalone installation of CocoaPods.

CocoaPods/Nanaimo 34

A native ruby implementation of ASCII plist parsing & serializing

endocrimes/Bees 18

A Swarm/Foursquare API Client written in Elixir

a2/passcards-swift-old 4

301 Moved permanently

endocrimes/AssetGen 3

A small Swift command line tool for generating Swift code for Asset Catalog resources

issue commentkubernetes/kubernetes

[Failing Test][sig-node] node-kubelet-serial-containerd

/triage accepted /priority important-soon /assign

RinkiyaKeDad

comment created time in 3 days

pull request commentkubernetes/kubernetes

Add e2e test for recreating static pods

/cc

ehashman

comment created time in 6 days

delete branch endocrimes/kubernetes

delete branch : dani/device-driver-deflake

delete time in 6 days

pull request commentNixOS/nixpkgs

go_1_15: set knownVulnerabilities

We might want to avoid doing this to avoid it getting backported to stable? (and wait until we can cleanly remove it instead) - otherwise we’ll have a repeat of 1.14’s deprecation.

zowoq

comment created time in 6 days

pull request commentkubernetes/kubernetes

WIP: kubelet: Handle UID reuse in pod worker

/assign

smarterclayton

comment created time in 6 days

pull request commentkubernetes/kubernetes

[Failing Test] Fix GPU Device Driver test in kubelet-serial

/hold cancel

:shipit:

endocrimes

comment created time in 6 days

Pull request review commentkubernetes/kubernetes

Revert #104308 to bring back LockContention tests

+// +build linux++/*+Copyright 2021 The Kubernetes Authors.++Licensed under the Apache License, Version 2.0 (the "License");+you may not use this file except in compliance with the License.+You may obtain a copy of the License at++    http://www.apache.org/licenses/LICENSE-2.0++Unless required by applicable law or agreed to in writing, software+distributed under the License is distributed on an "AS IS" BASIS,+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.+See the License for the specific language governing permissions and+limitations under the License.+*/++package e2enode++import (+	"time"++	"golang.org/x/sys/unix"++	"github.com/onsi/ginkgo"+	"github.com/onsi/gomega"+	"k8s.io/kubernetes/test/e2e/framework"+)++const contentionLockFile = "/var/run/kubelet.lock"++var _ = SIGDescribe("Lock contention [Slow] [Disruptive] [Serial] [NodeFeature:LockContention]", func() {

I think this might have to stay serial? - (given that it messes with kubelet health)

ipochi

comment created time in 7 days

PullRequestReviewEvent

pull request commentkubernetes/kubernetes

[Failing Test] Fix GPU Device Driver test in kubelet-serial

Leaving this until Monday afternoon (european time) and then I'll remove the hold if nobody gets to it before then

endocrimes

comment created time in 7 days

Pull request review commentkubernetes/community

sig-node: add doc with basic info for pause image updates

+# Updating Pause Image References++The Kubernetes `pause` image that is used to run a 'parent' container for all

I don't think it's worth it - linked docs about it tho

endocrimes

comment created time in 7 days

PullRequestReviewEvent

push eventendocrimes/community

Danielle Lancashire

commit sha 06fcf734d74705a8d725770a18fc5280e9e3c484

sig-node: add doc with basic info for pause image updates

view details

push time in 7 days

pull request commentkubernetes/kubernetes

update pause image references to use 3.6

@claudiubelu https://testgrid.k8s.io/sig-node-presubmits#pr-node-kubelet-serial - there hasn't been a clean run of that suite in as long as test grid has history - no need to keep re-testing.

claudiubelu

comment created time in 8 days

PR opened kubernetes/community

sig-node: add doc with basic info for pause image updates

This doc serves as a rough tracker of where we bump kubernetes Pause images, and how we try to coordinate them when new pause releases happen.

This is a first attempt to formalize roughly what happens in staggered PRs historically, and tries to keep things well tested while moving fairly quickly to unblock developers who may rely on new features for their testing (e.g Windows Server 2022 support in the latest release).

We may want to add a tracking issue template to this document after @kerthcet drafts the first issue.

Slack Discussion: https://kubernetes.slack.com/archives/C0BP8PW9G/p1631032292078800

/cc @dims /cc @SergeyKanzhelev /cc @BenTheElder /cc @kerthcet

+30 -0

0 comment

1 changed file

pr created time in 8 days

create barnchendocrimes/community

branch : dani/take-a-pause

created branch time in 8 days

PR opened kubernetes/community

sig-node-e2e: gcloud application-defaults are required

It wasn't obvious to me until reading the source code for the remote runner, and then looking up the GCP library docs that gcloud login wasn't sufficient to get things up and running.

This adds a quick note to the existing instructions for future folks who may want to run E2E's remotely.

+1 -0

0 comment

1 changed file

pr created time in 8 days

create barnchendocrimes/community

branch : dani/e2e-node

created branch time in 8 days

pull request commentkubernetes/kubernetes

Automated cherry pick of #104268: kubelet: fix sandbox creation error suppression when pods are

I'm not sure if this meets the requirements for a cherry pick - When the bug is encountered it's mostly a noise issue rather than one that can cause issues within the node or a cluster? (which doesn't fit in Loss of data/Memory corruption/Panic, crash, hang/Security)

rphillips

comment created time in 9 days

pull request commentkubernetes/kubernetes

update pause image references to use 3.6

/lgtm (kubelet changes and node-kubelet-serial state)

claudiubelu

comment created time in 9 days

pull request commentkubernetes/kubernetes

update pause image references to use 3.6

/lgtm (kubelet changes and node-kubelet-serial state)

claudiubelu

comment created time in 9 days

pull request commentkubernetes/kubernetes

update pause image references to use 3.6

@claudiubelu kubelet serial tests are known-flaky. Looks like no new or unexpected failures happened during the last run.

claudiubelu

comment created time in 9 days

pull request commentkubernetes/kubernetes

Revert #104308 to bring back LockContention tests

/triage accepted /priority important-longterm

ipochi

comment created time in 9 days

pull request commentkubernetes/kubernetes

Revert #104308 to bring back LockContention tests

@ipochi Unfortunately the easiest way to do that is by running them from your machine, which requires having a Google Cloud account configured, with a default application profile and project that won't interfere with any production environments you care about (gcloud auth application-default login)

and then e.g:

make test-e2e-node REMOTE=true FOCUS="LockContention" SKIP="" PARALLELISM=1 IMAGE_CONFIG_FILE="path/to/image-config.yaml" ZONE=us-central1-a
ipochi

comment created time in 9 days

pull request commentkubernetes/kubernetes

Support CRI `v1` and `v1alpha2` at the same time

logically speaking, this seems like the right approach over switching wholesale to 1.0, especially for future supportability and for deprecation cycles

/triage accepted

saschagrunert

comment created time in 9 days

Pull request review commentkubernetes/kubernetes

Ensure there is one running static pod with the same full name

 func (p *podWorkers) UpdatePod(options UpdatePodOptions) { 	var podUpdates chan podWork 	var exists bool 	if podUpdates, exists = p.podUpdates[uid]; !exists {+		fullname := kubecontainer.GetPodFullName(pod)+		if p.isConflictedPod(pod) {+			if pendingWork, exists := p.reservedWorkFullnames[fullname]; exists {+				pendingPod := pendingWork.Options.Pod+				if pendingPod.UID != uid {+					klog.V(4).InfoS("Mark an outdated pod as finished", "pod", klog.KObj(pod), "podUID", pod.UID)+					p.podSyncStatuses[pendingPod.UID].finished = true+				}+			}+			klog.InfoS("Reserve a podWork for the conflicted pod", "pod", klog.KObj(pod), "podUID", pod.UID)
			klog.InfoS("Reserving work to update the conflicted pod", "pod", klog.KObj(pod), "podUID", pod.UID)
gjkim42

comment created time in 9 days

Pull request review commentkubernetes/kubernetes

Ensure there is one running static pod with the same full name

 func (p *podWorkers) managePodLoop(podUpdates <-chan podWork) { 			klog.ErrorS(err, "Error syncing pod, skipping", "pod", klog.KObj(pod), "podUID", pod.UID)  		case update.WorkType == TerminatedPodWork:+			klog.V(4).InfoS("Processing pod event done", "pod", klog.KObj(pod), "podUID", pod.UID, "updateType", update.WorkType)+			if p.queueReservedWork(pod) {+				if start := update.Options.StartTime; !start.IsZero() {+					metrics.PodWorkerDuration.WithLabelValues("terminated").Observe(metrics.SinceInSeconds(start))+				}+				continue

Add a comment explaining why we maintain the loop here?

gjkim42

comment created time in 9 days

Pull request review commentkubernetes/kubernetes

Ensure there is one running static pod with the same full name

 func (p *podWorkers) acknowledgeTerminating(pod *v1.Pod) PodStatusFunc { 	return nil } +// queueReservedWork queues a reserved work having the same full name as pod+// and completes the previous pod. It returns true if there is a reserved work+// else returns false+func (p *podWorkers) queueReservedWork(pod *v1.Pod) bool {
func (p *podWorkers) queueWorkForConflictedStaticPod(pod *v1.Pod) bool {

This currently only applies to a small set of pods, lets avoid having a generic name that's easy to piggyback onto in the future without clear intent

gjkim42

comment created time in 9 days

Pull request review commentkubernetes/kubernetes

Ensure there is one running static pod with the same full name

 func (p *podWorkers) UpdatePod(options UpdatePodOptions) { 	} } +// isConflictedPod returns true when there is a known pod with the same+// fullname as pod+func (p *podWorkers) isConflictedPod(pod *v1.Pod) bool {
func (p *podWorkers) isConflictingStaticPod(pod *v1.Pod) bool {

Lets help future folks and reduce the scope of the name a little

gjkim42

comment created time in 9 days

Pull request review commentkubernetes/kubernetes

Ensure there is one running static pod with the same full name

 func (p *podWorkers) acknowledgeTerminating(pod *v1.Pod) PodStatusFunc { 	return nil } +// queueReservedWork queues a reserved work having the same full name as pod

Can you add the why here? - It won't necessarily be immediately obvious to future readers of this code.

gjkim42

comment created time in 9 days