A bug in systemd version 237-3ubuntu10.54 for the Ubuntu 18.04 Bionic Beaver long term support Linux distribution is to blame for several Microsoft Azure cloud services and products going down worldwide.
Systemd is a configuration and user processes management software for Linux, and Azure customers with Unattended-Updates enabled received a version which causes domain name system (DNS) resolution errors, Microsoft said.
Azure Kubernetes Service - AKS - and services that rely on it has been impacted in multiple regions due to the DNS resolution bug in the Ubuntu virtual machines.
Microsoft said Azure Container Apps, Database for PostgreSQL, and VMware Solution are affected worldwide.
Automated mitigation steps are under testing by Microsoft currently, to be applied across impacted AKS resources once it's validated.
The company has also removed the offending Ubuntu updates.
Rebooting impacted virtual machines so that they get new address leases with fresh DNS resolvers via the dynamic host configuration protocol (DHCP) is a potential workaround, Microsoft said.
A script shared by an AKS user could also be used as a workaround to avoid having to restart all nodes.
