Cases12 cases

7 to 11 of 12

1. IL-0443 A Repro Aurora MySQL reader restarted unexpectedly and writer insert latency rose about sevenfold for a day or more

   Repro2019

   Symptoms

   A reader of Repro's Aurora MySQL cluster restarted for no known reason, and right after that the writer's insert latency rose to about seven times normal and writer performance stayed badly degraded for a long time. This happened more than once, twice within two weeks by 11 November 2019. While the writer was slow, the applications that process data collected by the Repro SDK fell behind, and the team could only watch until the writer recovered.

   High latencyReplication lagThroughput degradation

   Trigger

   Long-running queries on the reader; the author recalls that the cause was a query sent to the reader from Redash.

   Root cause

   Capacity exhaustion

   Component

   DatabaseDatabase replication

   Contributing factor

   Monitoring blind spot

   Remediation

   • Waited for writer performance to return, since the rollback segment release could not be sped up

   Action items

   • Add a monitor on the rollback segment, mysql.innodb.history_list_length in Datadog
   • Kill queries that run long enough to make the reader restart as soon as they are found
   • Set max_execution_time so queries are stopped after a fixed time

   Lessons

   Once the rollback segment has grown large, nothing speeds up its release across the cluster, so the way to avoid the latency is to keep transactions small and stop long-running queries before the rollback segment grows.

   View case
2. IL-0323 A RAID battery backup unit failure crashed db1075, the MariaDB primary of section s3, and every s3 wiki went read-only

   Wikimedia Foundation2019

   Symptoms

   At 18:40 UTC db1075, the primary database of section s3, stopped answering, and all s3 wikis went read-only because no writes could reach it. Reads kept working from the replicas. At 18:42:45 Icinga reported "PROBLEM - Host db1075 is DOWN: PING CRITICAL - Packet loss = 100%".

   Writes rejectedReplication lagHigh latency

   Trigger

   The RAID battery backup unit (BBU) of db1075, the s3 primary database host, failed and the whole host crashed.

   Root cause

   Hardware failure

   Component

   DatabaseDatabase replication

   Contributing factor

   Pages not deliveredHarmful manual action

   Detection

   Icinga reported db1075 HOST DOWN on IRC at 18:42:45, but that alert does not page; replica Slave IO alerts followed, users reported problems on #wikimedia-operations, and the first SMS page went out at 18:56. Volunteers and staff noticed the failure before the alerts did.

   Remediation

   • Power-cycled db1075 from the iDRAC after finding the BBU problem
   • Let MySQL finish InnoDB recovery on db1075
   • Removed read_only=ON from db1075 by hand
   • Ran puppet by hand a second time so pt-heartbeat started

   Action items

   • Implement or refactor a script that moves replicas when the primary is unavailable
   • Fix the MediaWiki heartbeat model so pt-heartbeat does not use a super-user, avoids a single point of failure and follows the real primary without depending on puppet
   • Decide what to do with the batch of hosts that already had BBU issues
   • Buy a new BBU for db1075
   • Remove db1075 from the primary role
   • Address MediaWiki spam while the database is read-only or the primary is unavailable
   • Make primary database hosts page on HOST DOWN
   • Track hardware errors better in Netbox

   Lessons

   A BBU failure should not crash a whole host, yet it had happened before on HP hosts. HOST DOWN on a primary database should page instead of only alerting on IRC. Starting pt-heartbeat through puppet races with MySQL startup, and the unnoticed failure kept lag reported and made the outage 8 minutes longer.

   View case
3. IL-0210 GitHub's MySQL clusters failed over to the US West Coast and the site served slow, stale, inconsistent data for 24 hours

   GitHub2018

   Symptoms

   At 22:54 UTC on 21 October 2018 GitHub's internal monitoring started alerting on many faults at once. By 23:02 UTC the first responders saw that the replication topologies of many MySQL clusters were in an unexpected state: the Orchestrator API showed topologies made only of servers in the US West Coast data center. When connectivity came back, the application tier had immediately started sending writes to the new West Coast primaries.

   Split brainFailover failureReplication lagHigh latency

   Trigger

   At 22:52 UTC on 21 October, routine maintenance to replace failing 100G optical equipment cut connectivity between GitHub's US East Coast network hub and its primary US East Coast data center for 43 seconds.

   Root cause

   DNS or network fault

   Component

   DatabaseDatabase failover managerDatabase replicationDatabase backups and snapshotsNIC, cable and switch port link

   Contributing factor

   Automation amplificationUnverified assumption

   Detection

   GitHub's internal monitoring raised alerts on numerous faults at 22:54 UTC, two minutes after the partition. By 23:02 UTC the first responders had found, by querying the Orchestrator API, that many database clusters had a topology with only US West Coast servers.

   Remediation

   • Locked the internal deployment tooling to stop further changes
   • Paused webhook delivery and GitHub Pages builds to stop jobs writing metadata
   • Restored the affected MySQL clusters from backups into the US East Coast data center and let them replicate from the West Coast
   • Found ways to restore directly from the West Coast to get around the throughput limits of downloading from off-site storage
   • Moved all database primaries back to the US East Coast data center
   • Spread read traffic over additional MySQL read replicas in the US East Coast public cloud so replication could catch up
   • Failed over to the original topology once the replicas were in sync
   • Processed the backlog of webhook events and Pages builds, pausing to raise an internal TTL after about 200,000 webhook payloads were dropped

   Action items

   • Reconcile the writes from the captured MySQL binary logs that were not replicated, automatically where possible and with user outreach otherwise
   • Make sure status updates can be published even when GitHub Pages builds are paused
   • Give more accurate public estimates of the time to repair
   • Configure Orchestrator so it cannot promote database primaries across regional boundaries
   • Move to a status reporting mechanism that shows the status of each component of the platform
   • Serve GitHub traffic from multiple data centers in an active/active/active design with N+1 redundancy at the facility level
   • Test assumptions about the system more proactively
   • Validate failure scenarios systematically, with investment in fault injection and chaos engineering tooling

   Lessons

   Orchestrator did what it was configured to do, but the application tier could not work with primaries on the other side of the country: leader election within one region is generally safe, while the sudden cross-country latency was a major contributing factor. The behaviour emerged from a partition larger than any GitHub had seen. Replication caught up along a power decay curve, not a straight line, so the linear estimates of recovery time were wrong. GitHub concluded that tighter operational controls and faster response are not enough on their own and that failure scenarios have to be validated before they happen.

   View case
4. IL-0430 Mackerel's Redis stopped answering during a replication setup, keepalived failed over by mistake, and Mackerel degraded

   Hatena2018

   Symptoms

   On 26 September 2018 at 10:51 JST the whole Mackerel monitoring service stopped working properly and its alive monitoring stopped. The application servers could not connect to Redis, which stores Mackerel's monitoring data, and could not return correct responses.

   Monitoring outageHigh latencyIntermittent timeouts

   Trigger

   Redis memory use had been growing, so operators set up replication in order to scale Redis up. Building the replication took a long time and Redis stopped answering for a while, and keepalived treated this as a node failure and started a failover nobody intended.

   Root cause

   Configuration change

   Component

   In-memory cacheDatabase replicationDatabase failover manager

   Contributing factor

   Automation amplificationOverlapping failures

   Remediation

   • Switched back to the correct Redis and briefly saw the application return
   • Restarted the application servers one after another, which did not improve the symptoms
   • Detected the organizations posting inappropriate metrics and blocked their requests
   • Adjusted the timeout interval of the application servers
   • Temporarily added capacity to Diamond, the TSDB
   • Increased the number of Kinesis shards and Redis Cluster shards
   • Scaled out the API application servers
   • Put Mackerel in maintenance mode, checked it internally and let outside requests back gradually

   Action items

   • Raise the number of keepalived health checks for Redis so a brief load increase does not trigger a failover
   • Scale Redis up
   • Add more application servers
   • Change the application to store only the monitoring data it needs in Redis, reducing Redis memory use
   • Build a way to detect inappropriate API requests quickly and show them on a dashboard
   • Build a function to block organizations making inappropriate requests quickly
   • Keep improving the accuracy of detecting inappropriate requests and consider adding features such as an API limit
   • Strengthen monitoring of the application servers' internal processing so problems are handled before they become outages

   Lessons

   Hatena did not find why recovery took so long, but expects that keeping failover from reacting to short load spikes and giving the application more performance headroom would prevent an outage this long. It also aims to keep problems contained to a small part of the service when a failure does happen.

   View case
5. IL-0296 Transloadit's us-east Redis cluster lost queued jobs and heartbeats, raising WORKER_JOB_ERROR errors for 81 customers

   Transloadit2016

   Symptoms

   Reports came in of more WORKER_JOB_ERROR errors, which tell Transloadit there is a communication problem between its Uploaders, the machines that receive files and orchestrate Assemblies, and its Drones, the machines that run encoding tasks. Many Assemblies became much slower because retries with exponential back-off kicked in, and a few could not be replayed.

   High error rateHigh latencyLoad imbalance

   Trigger

   A steep increase in HLS traffic, together with more Drones online processing the same volume faster, which put more load on the central Redis nodes.

   Root cause

   Capacity exhaustion

   Component

   In-memory cacheDatabase replication

   Contributing factor

   Monitoring blind spotUnverified assumptionHarmful manual action

   Detection

   Reports of an increased number of WORKER_JOB_ERROR errors.

   Remediation

   • Re-added lost messages and relaxed heartbeat monitoring as a stopgap
   • Upgraded Redis to the latest stable version
   • Doubled the Redis cluster to six nodes and then added three more, which made it worse
   • Brought the number of read replicas down to one, which cut traffic and failure rates
   • Capped how far Drones scale up in each region

   Action items

   • Add logging and alarms for approaching saturation
   • Find ways to decentralise the traffic between the Redis nodes and the Drones and Uploaders
   • Add sharding per Uploader on top of the sharding per region

   Lessons

   To handle more traffic Transloadit had to use fewer Redis nodes, not more: each added replica multiplied the replication traffic from the single leader that takes every write, so scaling up the number of nodes was the wrong call. Bandwidth had not been among the monitored Redis metrics.

   View case

By root cause7