On February 2, 2022, incorrect IPv6 routes were inadvertently advertised due to a conflicting networking configuration on a newly-installed rack of hosts. This resulted in a data center wide IPv6 networking outage at the London data center. Customer Linodes in the London data center were provided incorrect default routes and this led to a loss of IPv6 connectivity at the Linode level.
At approximately 11:18 UTC, we began to see failures in the provisioning of a newly-deployed rack of hosts within the London data center. Around 11:30 UTC, we began receiving reports from customers reporting outages on IPv6 addresses. At 13:00 UTC, the Networking team identified the IPv6 routing issue, and at 13:15 UTC, the newly-provisioned rack of hosts was taken offline. This corrected the routing issues, and IPv6 connectivity was restored at 13:40 UTC.
We have made improvements to our host deployment processes by creating safeguards that prevent the deployment of an improper networking configuration. We've also made improvements to our deployment processes that will allow us to more quickly identify potential problems on newly-provisioned hosts.
If you have any further questions for us, please open a Support ticket for assistance.