Please find below the RFO for the RS1 session flaps.
Tue 13:58 - Reloaded network settings on RS1 in stead of new RS2, this caused the interface to reset and IPv4 and IPv6 sessions to flap
Tue 14:05 - Received alert, started investigation
Tue 14:20 - All sessions restored
Tue 14:22 - Sent message to mailing list
Tue 17:30 - Discovered IPv4 link local address on interface
Tue 17:40 - Removed address, this caused BIRD to update its interface's primary address (unknown than) and IPv4 sessions to flap
Tue 17:43 - Noticed flaps, started investigation
Tue 17:58 - Sent message to mailing list
Tue 18:01 - All sessions restored
Wed 15:22 - Reloaded network settings on new RS2, this started IPv4 link local address selection on new RS2 which sent ARP requests for verification
Wed 15:22 - This in turn triggered RS1 to do the same which then assigned itself an additional IPv4 address, which again caused IPv4 sessions to flap
Wed 15:30 - Noticed flaps, started investigation on RS1
Wed 15:42 - All sessions restored
Wed 15:47 - Sent message to mailing list
Wed 16:57 - No cause found yet, decided to postpone RS2 maintenance because of unpredictability and sent another announcement
Wed 17:30 - Started packet trace investigation which in the end showed the IPv4 link local issue
To avoid this in the future we're planning a number of changes for which we'll send a separate announcement.
If you have any questions don't hesitate to contact us.