Yesterday I updated a couple of my hosts to the latest CentOS packages as well as the latest ovirt packages, in the 3.5 series.

All looks fine, but once every couple of hours a host becomes completely unresponsive, doesn't even ping. Engine takes care of this - fencing is done. Nothing to be found in the logs of the host itself, engine simply tells me host became unresponsive.

Nothing else in the infra has changed. The other hosts, still at CentOS 7.1.1503, are fine, only the updated hosts at 7.2.1511 have this issue. 

I suspect a network driver issue somewhere, has anyone had the same experience so far? I'm using Intel X540-AT2 10 gbit cards in all my nodes, setup with LACP bonding.
--