Systems rebooted on H1
h1susauxey , every susaux on the cornerstation, and perhaps every front end updated to 5.0.1 rebooted on its own about 7 minutes after the first boot.
Three susaux servers on the corner station were observed to reboot in quick succession in the order and timing they were started, meaning the failure was tightly timed.
Below is a graph of h1susauxh2 cpu usage during the restarts. The first peak is startup CPU usage after the first planned restart. CPU usage is normal for about 7 minutes, then goes flat at just below 1%, then the server restarts.
After a few minutes we manually restarted the server agin.
Note that the flat period in the autoshutdown is lower cpu usage than in an ordered shutdown. Logs exported to the bootserver do not show any of the usual shutdown messages. The last message is one of normal operation. This does not appear to be a normal systemd shutdown.