Hello José, [2][3] - Pods that are deployed using host networking will use the node IP address (e.g. DaemonSets). Since the node IP address is within the management network subnet, then the deployment would be setup with multi-netting of the same physical interface and therefore the IP address is selected from that interface. Regards, Matt On 2019-04-11, 8:26 AM, "Perez Carranza, Jose" <jose.perez.carranza@intel.com> wrote: Hi all I'm reviewing this patch [1] and on the description says "System pods will be active on both controllers" but I'm checking according to below list [3] that 'Tiller-deploy' POD is only running on controller-1 that is on stand-by, is this normal behavior? Also according to this patch [2] says that pod network should be pointing to "172.16.0.0/16", but on the list [3] I see only 'tiller-deploy' and 'calico-kube-controllers' PODs using IPs on that range, is that normal? 1 - https://review.openstack.org/#/c/587458/ 2 - https://review.openstack.org/#/c/587465/ 3- http://paste.openstack.org/raw/749175/ Regards, José _______________________________________________ Starlingx-discuss mailing list Starlingx-discuss@lists.starlingx.io http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss