Late Saturday, September 13, an issue arose with our automatic DHCP support of registered desktop machines (wired network support of desktops/workstations without fixed IP addresses).
This appears to have affected devices with hostnames beginning with letters r through z, which have network registrations older than one year. Those devices would have lost their DHCP lease, and so would have lost their network connection, sometime early Sunday.
We have restored a known-good version of the DHCP support file for these hosts, and they should generally come back online with no intervention needed. Some devices may need to be rebooted, if they don’t come back online automatically.
Staff are continuing to investigate the root cause of the issue.
UPDATE
[2025-09-15 | larsrohr ]
While client support was fixed at about 10am, staff continued to investigate the root cause. Suffice it to say, a tab-vs-space issue in some perl code edited last October, coupled with a truncated sqllite database copy on Saturday, caused a failure to populate a file with all appropriate entries. That code has been fixed, and backend processes are again working as expected.
Resolved as of 2025-09-15 3pm