Cases16 cases

11 to 15 of 16

1. IL-0196 Adding a second Redis replica and lowering memory limits in one RedisFailover change restarted the only Redis pod that held data, and Sentinel promoted the empty new replica, erasing the data

   Palark2020

   Symptoms

   Right after an updated RedisFailover resource was rolled out, the large Redis database lost all its data and the service went down; the data had to be restored from a backup taken just before the change.

   Data lossService outage

   Trigger

   A RedisFailover change that raised the number of Redis replicas from one to two and, in the same rollout, lowered the containers' memory requests and limits from 22 GB to 19 GB.

   Root cause

   Software defect

   Component

   DatabasePersistent volumes and stateful setsContainer memory limit

   Contributing factor

   No staged rolloutSingle point of failureUnverified assumption

   Remediation

   • restored the Redis data from the backup prepared before the rollout

   Action items

   • Report the readiness probe flaw to the Redis Operator project and test the developers' fix
   • Reconsider the use of the Spotahome Redis Operator
   • Clean up the Redis databases and fix the application code that created the excess keys
   • Consider Redis built-in clustering and sharding

   Lessons

   Changing the replica count and the pod resources in one rollout turned a seemingly harmless change into a rolling restart of the only pod with data while the new replica was still empty. A readiness probe that only checks that Redis answers a ping does not tell whether a replica has finished syncing. Operators for critical stateful components need to be evaluated for maturity before production use, and the dump.rdb file should be backed up before any change to the Redis cluster.

   View case
2. IL-0193 A GCP persistent disk degradation took down a Cassandra StatefulSet whose OrderedReady policy blocked recovery, and Metrictank could not reconnect to the recreated pods, causing a 23-hour Graphite query outage

   Grafana Labs2019

   Symptoms

   An alert fired because some Metrictank instances could not drain their write queue, which meant they could not write to Cassandra. Google had not yet published an incident, but a team member's Discord server, also hosted on GCP, was down, which pointed at GCP; Google confirmed an SSD persistent disk incident about 90 minutes after it began.

   High error rateNode NotReadyPods PendingService outage

   Trigger

   A Google Cloud incident in which SSD persistent disks performed so poorly that many I/O operations failed for an extended period.

   Root cause

   Cloud provider outage

   Component

   Persistent volumes and stateful setsDatabaseNode root disk

   Contributing factor

   Overlapping failuresUnpatched known bugUnverified assumption

   Detection

   An alert that Metrictank write instances could not flush their write queue to Cassandra. Error rates alone rose too slowly to notice quickly, because most queries were served from memory.

   Remediation

   • Added compute nodes hoping new nodes could attach the disks, which did not work
   • Deleted and recreated the Cassandra StatefulSet with an updated pod management policy, so all pods but one started
   • Raised Kafka topic retention to avoid data loss during the long downtime
   • Temporarily added nodes so Metrictank restarts were not slowed by CPU starvation
   • Restarted all Metrictank instances, which took more than 20 hours, so they connected to the new Cassandra IPs

   Action items

   • Re-instantiate the gocql connection in Metrictank when it is detected dead, so it reconnects even when all Cassandra IPs change
   • Fix the Metrictank write-queue deadlock by updating an internal data structure with atomics instead of a lock

   Lessons

   An OrderedReady StatefulSet cannot recover past a single pod whose volume will not attach; when every pod is down, the ordering that protects normal start-up blocks the whole cluster. A client that learns peer addresses only from live peers needs a way to go back to the service name when all peers are replaced at once. A long-tolerated bug with a config workaround can surface at the worst time; this one turned out to have a simple fix. Writing incoming data to a Kafka write-ahead log first meant all data could still be ingested after recovery, and Kafka on non-SSD disks escaped the disk incident. With both Metrictank bugs fixed, the team expects a similar event to take less than 3 hours to recover from instead of 23.

   View case
3. IL-0518 gitlab.freedesktop.org on GKE ran short of volume space and a planned 30-minute storage upgrade became a 4-hour outage

   freedesktop.org2018

   Symptoms

   Shortly after 9am UTC on 17 July 2018, the disk space of the GitLab cluster behind gitlab.freedesktop.org was critical, and some API requests were failing because of the shortage. The administrator first learned of the failure when people mentioned it on IRC.

   Persistent volume fullHigh error rateService outage

   Trigger

   Many recently created forks of Mesa, each about 2.1GB, filled the disk space of the GitLab cluster.

   Root cause

   Capacity exhaustion

   Component

   Persistent volumes and stateful setsCloud block storageHelm

   Contributing factor

   Unverified assumptionUnrepresentative stagingMonitoring blind spotChange during peak hours

   Detection

   People mentioned the failure on IRC. Monitoring and alerting on disk usage and API error rate had not been set up yet.

   Remediation

   • Dropped all inbound traffic to GitLab at the firewall except from the administrator's IP
   • Took a manual backup by streaming tar over SSH to an encrypted partition on a local machine
   • Snapshotted all GCE disks behind the PersistentVolumes
   • Copied the data from the old disk snapshots to new enlarged volumes through a temporary GCE VM
   • Re-did the copy into new empty volumes with the correct directory structure and checked repositories, users and workflows
   • Restored public access to GitLab

   Action items

   • Set up monitoring, logging and alerts
   • Find ways to communicate outages much better; this one was announced only on IRC, and the lists were not told when it dragged on
   • Use the Cloud console clients instead of the slow Cloud Console web UI
   • Add more GitLab administrators, ideally someone familiar with Kubernetes deployments
   • Move from the single Omnibus container to GitLab's cloud native deployment with one container per service
   • Move storage to resizeable volumes or to Google Cloud Storage

   Lessons

   The backup step alone used up the outage window; backups had grown along with the disk problem, and a test run before announcing the window could have measured it. The changes had been tested on a scratch cluster, but not exactly as run, and tests of steps such as disk snapshots could have been more realistic. A single GitLab administrator who was new to Kubernetes was a single point of failure. Keeping at least three copies of the data from the first destructive step meant no data was lost.

   View case
4. IL-0402 mybinder.org launches dropped to zero for hours because a JupyterHub pod on a lost GKE node kept its hub database PVC

   Binder Project2018

   Symptoms

   On 13 March 2018 at 07:28 CET the mybinder.org team noticed that JupyterHub had not been spawning new users and that the launch success rate on the Grafana dashboard had been at zero for several hours. Many pods were in the states "Unknown" and "NodeLost", on at least the nodes `10.128.0.7` and `10.128.0.5`.

   Service outageNode NotReadyPods stuck in TerminatingPods stuck in ContainerCreating

   Root cause

   unknown

   Component

   Persistent volumes and stateful setsNode pool

   Contributing factor

   Pages not delivered

   Detection

   The team noticed only after several hours, when the launch success rate on the Grafana dashboard had been at zero for a long time. No Stackdriver alert email arrived in time, because those emails were marked as spam and reached the team hours later.

   Remediation

   • Cordoned the bad nodes
   • Deleted the hub pod in Unknown state and the hub pod stuck in ContainerCreating
   • Reset node gke-prod-a-ssd-pool-32-134a959a-n2sk from the GCP web interface to force the PVC to be released
   • Force-deleted every pod in Unknown or NodeLost state with kubectl delete --grace-period=0 --force

   Action items

   • Investigate why the outage started, through Stackdriver logs and Google Cloud incidents
   • Set up an alerting service that does not depend on Stackdriver
   • Document how to check whether a PVC has not been released
   • Document how to restart a node by hand when commands are not working
   • Document how to delete pods that are in Unknown state

   Lessons

   The error went unnoticed for a long time because Stackdriver alert emails were marked as spam and arrived hours late. Once the problem was found, each step was logged and the steps were made more forceful as needed, so the hub was down only briefly after detection.

   View case
5. IL-0072 High etcd write latency made the API servers return stale data, and the production cluster flipped between two views of its state

   Nordstrom2017

   Symptoms

   A user reported that pods were starting up and then vanishing. When the cluster operators looked, kubectl get pods and kubectl get nodes returned two different sets of data that alternated back and forth, as if there were two opinions of the cluster's state.

   Stale or inconsistent API readsStale endpoint routingService outage

   Trigger

   etcd write latency rose far above the healthy single-digit milliseconds; the speakers say trouble starts once the database sync time crosses double digits or several hundred milliseconds.

   Root cause

   Software defect

   Component

   etcdKubernetes API serverIngress controllerPersistent volumes and stateful sets

   Contributing factor

   Unreviewed defaults

   Detection

   A user reported pods starting up and vanishing.

   Remediation

   • restored service on a replacement cluster
   • freed persistent volumes from the old cluster one by one so they could bind in the new cluster
   • updated DNS records that pointed at the old cluster's load balancer names

   Action items

   • Make the Kubernetes API server request quorum reads from etcd

   Lessons

   etcd is extremely sensitive to write latency; once the database sync time crosses double digits of milliseconds, stale reads or badly degraded performance follow. Without quorum reads, a consistent store can still hand out stale data. Documentation for a fast-moving system changes, so read it again regularly. Moving workloads to a new cluster is hard when they hold persistent volumes or when DNS records point at generated load balancer names.

   View case

By root cause9