11-18-2021 07:25 AM
I post here because we have had an issue with an HA pair when we was upgrading the ACI.
To explain the architecture, the HA pair are in two different datacenter and have LAN1 and HA connected to one switch. LAN2 (configured like LAN1 redundancy) and MGMT are connected to a different switch.
So during the upgrade, the LAN1 and HA port of the active node was interrupted and the passive tried to become active but the VIP of HA did not appear until we are disabling the rogue ep control on ACI.
The rogue ep control was in default setting (rogue ep control intervalle : 60, rogue ep detection multiplication factor : 4, hold interval (1800).
Have you got the same problem or have you some advice ?
12-15-2021 07:49 AM
Hi i have had this problem on ACI
With another client we clould not reproduce this on the 8.5.2 nios firmware and ACI 4.2.6(d) release train. The problem i posted was at antother client that was still on 8.4.x nios version and ACI 3.2.6 release.
In both cases we don't use rouge endpoint control.
For other best practices on ACI this is also a good site :
https://unofficialaciguide.com/ for bestpractices in settings on Muti-pod fabric e.t.c