Cases7 cases

4 to 7 of 7

1. IL-0240 Cloudflare API requests failed and the dashboard slowed for hours after the authentication database lost its replicas

   Cloudflare2020

   Symptoms

   On 2 November 2020 the Cloudflare API and dashboard were degraded for six hours and 33 minutes. The success rate of API queries dropped at times to 75%, and the dashboard was up to 80 times slower than normal, while Cloudflare's edge network kept working.

   High error rateHigh latencyDatabase overloadWrites rejectedPacket loss

   Trigger

   At 14:43 UTC a network switch in a control plane data center partly failed: it kept LACP and BGP running but stopped forwarding some packets, and it recovered on its own six minutes later.

   Root cause

   DNS or network fault

   Component

   NIC, cable and switch port linkCoordination and key-value storeDatabase failover managerDatabaseDatabase replication

   Contributing factor

   Automation amplificationNo headroomHarmful manual action

   Detection

   At 14:43 UTC alerts fired because the misbehaving switch was unreachable to pings. Engineers then saw that the primary authentication database was overwhelmed.

   Remediation

   • Shed load from the primary authentication database by slowing SSL certificate pushes to the edge, emails and other features
   • Moved API traffic that could use read replicas to the secondary data center by hand
   • Sent authentication calls fully back to the primary data center when the dashboard got worse
   • Rebuilt the first database replica, which put itself back into service at 21:20 UTC and returned performance to normal
   • Ramped the turned-down services back up so asynchronous processing could catch up

   Action items

   • Remove the cluster management limitation that requires rebuilding all replicas after a primary is promoted
   • Redesign the user sessions system so dashboard traffic can be steered between data centers
   • Change the auto-remediation configuration so a database replica is not promoted so quickly
   • Tune each system involved to be more resilient to failures of its dependencies

   Lessons

   Every system in the chain had redundancy and none failed completely; each only degraded, which made the chain of events much harder to model and foresee. Automating replica promotion so that it could trigger on a minute's notice had been a point of pride, but for at least one database a fast promotion can do more harm than the failure it answers. Most general-purpose cluster managers use Raft or Paxos rather than Byzantine fault tolerant protocols, accepting a rare failure mode in exchange for a protocol that is simpler to implement correctly and debug.

   View case
2. IL-0218 GitLab.com's primary PostgreSQL data was wiped by mistake and, with no working backup, the site was down about 18 hours

   GitLab2017

   Symptoms

   From around 19:00 UTC on 31 January 2017, load on GitLab.com's PostgreSQL database rose, which GitLab suspected was spam, and many users could not post comments on issues and merge requests. Around 23:00 UTC the replication of the PostgreSQL secondary fell behind and then failed, because WAL segments it still needed had already been removed from the primary.

   Database overloadReplication lagData lossService outage

   Trigger

   While trying to restore replication to the secondary, an engineer ran the removal of the PostgreSQL data directory on the primary, db1.cluster.gitlab.com, believing it was the secondary.

   Root cause

   Operator error

   Component

   DatabaseDatabase replicationDatabase backups and snapshotsCloud block storage

   Contributing factor

   Untested backupSingle point of failureNo runbookPages not deliveredUnverified assumptionUnreviewed defaults

   Detection

   The engineer who ran the deletion on the primary noticed the mistake and stopped it a second or two later.

   Remediation

   • Raised max_wal_senders from 3 to 32 and lowered max_connections from 8000 to 2000 so PostgreSQL would restart
   • Stopped the deletion on the primary a second or two after noticing it
   • Restored the database from the LVM snapshot taken about 6 hours before the outage by copying the PostgreSQL data directory from staging
   • Restored the webhooks from a separate database built from the snapshot and imported them as a SQL dump
   • Incremented all database sequences by 100,000 so IDs used before the outage could not be reused
   • Re-enabled GitLab.com gradually
   • Rebuilt projects created after 17:20 UTC in basic form from their Git repositories on the NFS cluster

   Action items

   • Change the shell prompt (PS1) on all hosts to show more clearly which host and environment is in use
   • Monitor backups with Prometheus
   • Set PostgreSQL's max_connections to a sane value
   • Investigate point-in-time recovery and continuous archiving for PostgreSQL
   • Take LVM snapshots of the production databases every hour
   • Take Azure disk snapshots of the production databases
   • Move staging to the ARM environment
   • Recover the production replicas
   • Test recovery of PostgreSQL database backups automatically
   • Improve the PostgreSQL replication documentation and runbooks
   • Investigate pgbarman for PostgreSQL backups
   • Investigate WAL-E for database backup and realtime replication
   • Build a streaming database restore
   • Assign an owner for data durability
   • Set up multiple secondaries and balance load among them with pgpool-II
   • Provide a tool for developers to run and revert Rails migrations on staging
   • Build owned recovery procedures for the rest of the GitLab.com infrastructure, not only the database
   • Publish a backup monitoring dashboard and add more backup data to it
   • Improve the abuse reporting and response system so removal for spam does not hard delete
   • Redact engineers' names in future public incident documents

   Lessons

   GitLab split the incident into two problems: GitLab.com went down because the primary's data directory was removed by mistake during a manual, undocumented replication rebuild, and recovery took over 18 hours because the backups had failed silently, disk snapshots were not enabled and the only copy sat on slow staging disks. The backup procedure was never tested because no one owned it. GitLab chose to focus on fast, regularly practised recovery and on making it obvious which host an engineer is on, rather than on blocking commands, since a safer rm only protects against one of many ways to lose data.

   View case
3. IL-0215 Twilio's billing Redis master lost all balances and went read-only, and auto-recharge charged customer cards repeatedly

   Twilio2013

   Symptoms

   At 2:39 AM PDT (9:39 UTC) on 18 July 2013, services that relied on the redis-master of Twilio's billing system began to fail because the load on its host had become extreme. At 2:42 AM PDT monitoring alerted the on-call engineers to a failure in the Redis cluster.

   Database overloadData lossWrites rejected

   Trigger

   At 1:35 AM PDT on 18 July a loss of network connectivity disconnected all billing redis-slaves from the redis-master at once, and all of them requested a full synchronization at the same time. On-call engineers then restarted the overloaded redis-master.

   Root cause

   Configuration change

   Component

   In-memory cacheDatabase replication

   Contributing factor

   Retry stormHarmful manual actionUnsafe failure behaviour

   Detection

   At 2:42 AM PDT monitoring alerted the on-call engineers to a failure in the Redis cluster. At 3:28 AM PDT the billing system monitoring reported the anomalous charges, and at 12:36 PM PDT monitoring detected the recurrence.

   Remediation

   • Restarted the redis-master to relieve the high load, which loaded a wrong configuration
   • Took the billing system offline to stop further credit card charges
   • Restored service to all suspended accounts
   • Restored the lost account balances from the independent double-bookkeeping record in the relational datastore
   • Took billing offline again after the recurrence and reactivated the newly suspended accounts
   • Refunded all erroneous credit card charges
   • Gave all affected accounts a credit of 10% of their last 30 days of spend
   • Replaced the original Redis cluster and corrected the redis-master configuration
   • Activated the billing system progressively for groups of customer accounts

   Action items

   • Disable Redis restarts on the redis-master and recover the master by pivoting a slave instead
   • Add fail-safes so the billing system neither suspends accounts nor charges cards when balances are missing or cannot be written
   • Validate the billing system against the double-bookkeeping databases in real time

   Lessons

   Losing the in-flight balances and the ability to write balances at the same time exposed a flaw in the auto-recharge system: it failed in a dangerous way, charging cards and suspending accounts instead of stopping.

   View case
4. IL-0303 GitHub's MySQL master failed over to a node with a cold buffer pool during a schema migration and GitHub.com went down

   GitHub2012

   Symptoms

   On Monday, September 10th, 2012, during a schema migration, the health checks on GitHub's MySQL master failed and Percona Replication Manager moved the 'active' role to another node of the new three-node cluster. That node's InnoDB buffer pool was cold and it performed poorly, its health checks failed in turn, and the role failed back to the original server. GitHub.com went down.

   Failover failureDatabase overloadHigh latencyReplication lagSplit brainService outage

   Trigger

   A schema migration on Monday, September 10th put more load on the master database than earlier migrations had, and Percona Replication Manager's health checks on the master failed.

   Root cause

   Software defect

   Component

   DatabaseDatabase failover managerDatabase replication

   Contributing factor

   Automation amplificationDependence on a warm cacheHarmful manual actionOverlapping failures

   Detection

   On the morning of September 11th, a developer told the operations team that the 'standby' node was returning incorrect query results.

   Remediation

   • Enabled Pacemaker's maintenance-mode so no health checks or automatic failovers ran, and let the buffer pool warm
   • Powered off the out-of-date node 'c' to end the partition
   • Stopped Pacemaker and Heartbeat on nodes 'a' and 'b' and started them on node 'a' only
   • Started MySQL on node 'a' in the active role
   • Removed the events leaked to wrong dashboards and audited the wrongly routed repositories
   • Contacted the owners of the exposed private repositories

   Action items

   • Change the Pacemaker configuration so the active database role fails over only when an operations engineer starts it
   • Find a way to warm the InnoDB buffer pool of slave databases so failovers do not hurt performance
   • Audit the Pacemaker and Heartbeat stack, focusing on the code path behind the segfault
   • Run hardware tests on the server where the segfault happened

   Lessons

   GitHub concluded that automated failover is not the right strategy for its primary production database: for every failover in this incident, any member of the operations team would have said it should not happen. Failovers that must happen, in an emergency or during maintenance, should not hurt performance.

   View case

By root cause7