Cases19 cases

14 to 18 of 19

1. IL-0014 A GKE upgrade to a withdrawn version stalled on slow drains and a stuck node pool, then broke an ingress and confused the HPA

   loveholidays2019

   Symptoms

   On the evening of a long-overdue upgrade of a loveholidays production GKE cluster from 1.10.X to 1.12.5-gke.10, planned to take 2 to 2.5 hours, the master upgrade and the first two 7-node NodePools went smoothly. After that the NodePool upgrades slowed to a crawl. Each node is cordoned, drained and upgraded in turn. Even a 14-node cluster at about 5 minutes per node took over an hour, and with about 100 more nodes to go, some of them taking over 15 minutes to drain, the sequential drains were too slow to finish at a reasonable hour.

   Upgrade stallPods PendingScale-up failureService outageAutoscaler misbehaviour

   Trigger

   A manual upgrade of the GKE masters and NodePools from 1.10.X to 1.12.5-gke.10, two versions at once, on 19 March 2019; GKE's release notes of 14 March had already said that version was no longer available for new clusters or master upgrades because masters showed elevated error rates, but it could still be selected.

   Root cause

   Cloud provider outage

   Component

   GKE control planeNode poolCluster autoscalerIngress controllerHorizontal pod autoscaler

   Contributing factor

   Unrepresentative stagingOverlapping failuresIgnored early signal

   Detection

   The slow drains, zombie nodes and stuck NodePool were seen by the engineer running the upgrade; the broken Ingress was noticed after the upgrade and explained by Google support; the HPA problem was noticed the next day.

   Remediation

   • drained the remaining NodePool nodes ahead of the upgrade with kubectl drain --ignore-daemonsets --delete-local-data
   • deleted the zombie instances from the managed instance group with gcloud compute instance-groups managed delete-instances
   • cancelled the stuck UPGRADE_NODES operation with gcloud beta container operations cancel
   • used kubectl port-forward in place of the broken internal Ingresses
   • scaled down the older of two active ReplicaSets, or triggered a dummy deployment, to bring the HPA back

   Action items

   • Alert when a HorizontalPodAutoscaler is at 70% and at 100% of its capacity
   • Alert when a deployment has had two active ReplicaSets for 30 minutes and for 60 minutes
   • Re-evaluate termination grace periods that were set to arbitrary large values
   • Read the GKE release notes before every cluster upgrade
   • Use auto-upgrade in non-production environments, and manual cluster upgrades with automatic node upgrades in production
   • Keep environments as identical as possible, including resource names

   Lessons

   Read the release notes before upgrading: the target version had been withdrawn five days earlier. Drain time depends on pod configuration, so emptyDir volumes, long grace periods and missing disruption budgets turn a NodePool upgrade into hours. Pre-production only catches problems like the ingress bug if it matches production down to resource names. Auto-upgrades in non-production surface problems early, and a mix of manual cluster upgrades and automatic node upgrades in production keeps clusters current without much effort.

   View case
2. IL-0466 Prow scheduled test job pods onto its service cluster, leaving hundreds Pending and starving Prow's own services

   Kubernetes project2019

   Symptoms

   On 8 February 2019 the Kubernetes project's Prow instance at prow.k8s.io became very slow: its controllers took about five minutes to respond to the hook, bot commands such as the cat command stopped working, and the prow.k8s.io site was hard to reach. GitHub status was normal, and the Prow service nodes were found running above 100% of CPU and memory.

   Pods PendingSustained high node CPUHigh latency

   Trigger

   At 13:20 PST an auto-generated bump PR for Prow was merged and deployed, and Prow started scheduling workloads onto the service cluster.

   Root cause

   Faulty deployment

   Component

   CI system

   Contributing factor

   Unrepresentative stagingNo resource requests or limitsLimited debug access

   Detection

   While working on a PR, a maintainer noticed the bot's cat command failing, found GitHub status normal, learned on Slack that the Prow nodes were above 100% CPU and memory, and then found 600 pending pods in the test-pods namespace.

   Remediation

   • Rolled Prow back to v20190206-b01b288 and cleaned up the pods and ProwJobs in the service cluster
   • Applied the revert by hand with kubectl apply after the deployment permission was missing
   • Sent a fix, PR 11209, so the client no longer overwrites a provided literal default client
   • Rolled Prow back to v20190125-2aca69d, from before the bug was introduced
   • Cleaned up the wrongly scheduled pods and ProwJobs again and confirmed Prow was healthy

   Action items

   • Find out why the bug did not show deterministically
   • Be more careful with sweeping Prow changes, with more end-to-end tests or a real staging Prow

   Lessons

   The behaviour was not deterministic, and the change was deployed on a Friday afternoon. Prow instances are set up differently, for example in how build clusters are defined, so another project's Prow running a version fine does not show it is safe for the Kubernetes Prow, which needs a staging environment that looks like it.

   View case
3. IL-0504 mybinder.org launches failed for about five hours as undeleted user pods piled up and filled the cluster

   Binder Project2018

   Symptoms

   On Saturday 8 July 2018 at about 18:30 CEST, the launch success rate of mybinder.org dropped to zero and stayed there. Suspicious behaviour was first reported on Gitter at 22:48, more than four hours later.

   Service outagePods PendingStale or inconsistent API reads

   Root cause

   Software defect

   Component

   JupyterHub

   Contributing factor

   No alert on limit

   Detection

   Suspicious behaviour was reported on Gitter at 22:48 CEST, more than four hours after launch success dropped to zero; no downtime notifications were set up.

   Remediation

   • Deleted user pods older than four hours with the delete-pods.py script
   • Culled pods of the most-used images, such as ipython-in-depth, more aggressively
   • Restarted the Hub pod after about 600 pods were culled, which brought launches back to normal
   • Cordoned some of the extra nodes so the autoscaler could reclaim them
   • Ran manual max-age culling the next morning, letting the cluster scale back to 3 nodes

   Action items

   • Set up notifications of downtime
   • Automate cluster scale-down
   • Document the downtime recovery steps so every team member knows them
   • Fix KubeSpawner's failure to delete pods when the pod reflector fails
   • Make KubeSpawner more robust to pod reflector failure
   • Refactor the event reflector into a single reflector instead of one per pod
   • Detect a likely unhealthy Hub state and restart the Hub without manual intervention

   Lessons

   The delete-pods.py script was useful for bulk-deleting old pods, and once the problem was noticed the cluster was healthy within a few minutes. The outage lasted five hours before anyone noticed because no downtime notifications existed, and scaling the cluster back down still took several manual steps of cordoning and draining nodes.

   View case
4. IL-0038 kubelet rollout hit a Kubernetes bug and left every node in a cluster NotReady, triggering mass pod eviction

   Oath2018

   Symptoms

   During a rolling update of a kubelet change in one of Oath's on-premises production clusters, nodes started going NotReady, and soon all nodes in the cluster were NotReady.

   Node NotReadyPod evictionPods Pending

   Trigger

   A change to kubelet was rolled out across the nodes of one cluster; on restart, kubelet cleans up transient cgroup files, and a Kubernetes bug in that step left kubelet stuck instead of starting.

   Root cause

   Software defect

   Component

   kubeletKubernetes controller manager

   Contributing factor

   No health probesExcessive blast radius

   Detection

   Nodes reported NotReady and the eviction graphs climbed while the kubelet rollout was in progress.

   Remediation

   • failed the affected data center over so users were routed to the next available data center

   Action items

   • Run a functional test of kubelet after each upgrade batch and certify it before the next 30% of nodes are updated

   Lessons

   Check kubelet after upgrading it before moving to the next batch; a simple curl against the kubelet API would have caught this. The node controller's eviction throttling and full disruption mode, which the speakers note is not well documented, kept a cluster-wide kubelet failure from evicting every pod.

   View case
5. IL-0040 Parallel Helm installs saturated small gp2 root volumes on a kops cluster and left all three workers NotReady with the kubelet or Docker down

   Saltside2017

   Symptoms

   While running the CI process on a local machine to test an application Helm chart, the SRE lead watched the pods with watch kubectl get pods. All pods reached Running, and then some went back to Pending. Two hours later the problem had grown and some pods showed the phase Unknown.

   Pods PendingImage pull failuresNode NotReadyService outage

   Trigger

   A CI run installed the application's Helm charts for all four markets (Efritin.com, Tonaton.com, Ikman.lk and Bikroy.com) on the nonproduction cluster at the same time.

   Root cause

   Capacity exhaustion

   Component

   Node root diskkubeletContainer runtimeHelm

   Contributing factor

   Unreviewed defaultsMonitoring blind spotIgnored early signalSingle point of failure

   Detection

   The SRE lead saw pods go from Running back to Pending while watching kubectl get pods during the CI run; no monitoring raised an alert, even when all three workers were not ready.

   Remediation

   • terminated the three worker nodes and let the auto scaling group replace them

   Action items

   • Fix the Datadog integration for the nonproduction AWS account and repeat the installs with telemetry
   • Increase the worker root EBS volumes from 20GB to 120GB
   • Keep common base images pulled on every node with a DaemonSet until kops supports user hooks
   • Configure different allocation limits for the different layers in Kubernetes
   • Discuss a highly available Tiller with the Helm team
   • Find a way to throttle the number of newly created pods

   Lessons

   An ordinary deploy of the application could collapse the cluster, because the default 20GB gp2 root volumes could not keep up with parallel image pulls and container starts. Without node metrics and pod phase metrics, three NotReady workers raised no alert. A single Tiller replica can block the deploy that would fix an outage.

   View case

By root cause6