this issue has been closed. See "Milan DHCP Network Issue" incident to get further information on the resolution.
Jan 10, 11:22 CET
The situation is stable now. All DHCP agents have been restarted and work properly. We'll investigate the root cause of the problem.
Dec 29, 18:01 CET
We are experiencing issues with DHCP server on Milano, and we're investigating why after updating to Openstack Kilo some instances may lose connectivity.
We 'll keep you posted for updates. We apologise for the inconvenience.
Dec 29, 17:13 CET
In progress -
Scheduled maintenance is currently in progress. We will provide updates as necessary.
Dec 28, 09:00 CET
step 3 is the most sensitive upgrade overall: Networking (Neutron).
We’ll upgrade APIs, all L3 agents (including VPNaaS, LBaaS, FWaaS, DHCP and Internet Gateway) and L2 agent (virtual switches on compute nodes).
A downtime around a couple of minutes per tenant is expected while rebooting all virtual interfaces on Internet Gateways, plus a 15/20-seconds glitch on each compute node while re-creating L2 interfaces in Open vSwitch.
Therefore we’ll carry out this maintenance during nighttime to reduce possible drawbacks.
Order of execution will be: NL-AMS1, DE-FRA1, IT-MIL1, one region per day.
Updates will be available during operations.
Dec 7, 19:13 CET