Time: 22.07.2021, 8 am
A bug in the software Systemd, a widely used core component on Linux systems, can lead to failures especially on services that use programs to ensure high availability like Corosync, keepalived, pacemaker or heartbeat.
Both the failure of the service https://jupyter-cloud.gwdg.de on 21.07.2021 between 6:10am and 11:10am and today’s failure of https://meet.gwdg.de are caused by this bug. After the bug occurs, a restart of the affected HA service is sufficient to restore the servers high availability. As a countermeasure, automatic updates of the systemd package will be disabled on potentially affected servers for the time being.