I tried multiple permutations within Bright Cluster Manager but the first interface always is down after boot. To make it clear, let's say I have ens10p0 and ens10p1 interfaces as members of the bond, and pxe boots from ens10p0 . After boot, ens10p0 is always in a DOWN state.
As a temporary workaround I just made a small script into rc.local that ifups all the members of the bond. No need to change any configuration, because Bright Cluster Manager is creating the ifcfg files properly. It just happens that the iface gets down somewhere on boot.
Tried on different servers and all show the same behaviour, so I am pretty sure this is not a hardware problem.