Reproduced with RHEL-OSP6-Beta: openstack-packstack-puppet-2014.2-0.5.dev1316.g733aa73.el7ost.noarch openstack-packstack-2014.2-0.5.dev1316.g733aa73.el7ost.noarch +++ This bug was initially created as a clone of Bug #1163697 +++ Description of problem: ======================= This seems like a regression with packstack. I configured the answer file with provider ovs bridge, yet the bridge (named br-nodes) was tapped to the relevant NIC in the networker node but not in the compute nodes. The motivation for this test is to allow a provider IPv6 vlan network. In addition Used another NIC as a tunnel interface, to allow vxlan/gre tenant networks as well (for non-provider networks). snipped from answer file: CONFIG_NEUTRON_L3_EXT_BRIDGE=provider CONFIG_NEUTRON_L2_PLUGIN=ml2 CONFIG_NEUTRON_ML2_TYPE_DRIVERS=vxlan,vlan CONFIG_NEUTRON_ML2_TENANT_NETWORK_TYPES=vxlan,vlan CONFIG_NEUTRON_ML2_MECHANISM_DRIVERS=openvswitch CONFIG_NEUTRON_ML2_VLAN_RANGES=ipv6_vlan_range:60:66 CONFIG_NEUTRON_ML2_TUNNEL_ID_RANGES=1:1000 CONFIG_NEUTRON_ML2_VNI_RANGES=10:100 CONFIG_NEUTRON_L2_AGENT=openvswitch CONFIG_NEUTRON_OVS_TENANT_NETWORK_TYPE=vxlan,vlan CONFIG_NEUTRON_OVS_VLAN_RANGES=ipv6_vlan_range:60:66 CONFIG_NEUTRON_OVS_BRIDGE_MAPPINGS=ipv6_vlan_range:br-nodes CONFIG_NEUTRON_OVS_BRIDGE_IFACES=br-nodes:enp5s0f1 CONFIG_NEUTRON_OVS_TUNNEL_RANGES=1:1000 CONFIG_NEUTRON_OVS_TUNNEL_IF=enp4s0f1 Version-Release number of selected component (if applicable): ============================================================= RDO-Juno. openstack-packstack-2014.2-0.5.dev1316.g733aa73.el7.centos.noarch openstack-packstack-puppet-2014.2-0.5.dev1316.g733aa73.el7.centos.noarch How reproducible: ================= 1/1 Steps to Reproduce: =================== 1. Used the attached answer file with packstack. 2. post installation, check your compute nodes for the provider bridge ports: # ovs-vsctl show Actual results: =============== Bridge br-nodes Port phy-br-nodes Interface phy-br-nodes type: patch options: {peer=int-br-nodes} Port br-nodes Interface br-nodes type: internal Expected results: ================= Bridge br-nodes Port "enp5s0f1" Interface "enp5s0f1" Port phy-br-nodes Interface phy-br-nodes type: patch options: {peer=int-br-nodes} Port br-nodes Interface br-nodes type: internal Port "enp5s0f1\nbr-nodes" Interface "enp5s0f1\nbr-nodes" Additional Info: ================ Workround: ovs-vsctl add-port br-nodes enp5s0f1
Reopening. Got the exact same result with: openstack-packstack-2014.2-0.12.dev1370.ge46bee1.el7ost.noarch See logs attached
The original issue with packstack not tapping devices to the bridges when CONFIG_NEUTRON_ML2_TYPE_DRIVERS=vxlan,gre,vlan was set is fixed and works, the second issue with not tapping when vlan is missing from CONFIG_NEUTRON_ML2_TYPE_DRIVERS is not a bug.
I'm sorry my last comment was wrong, again: CONFIG_NEUTRON_ML2_TYPE_DRIVERS=vxlan,gre,vlan is comma separated list of network type driver entrypoints to be loaded from the neutron.ml2.type_drivers namespace. The original bug was about CONFIG_NEUTRON_ML2_TENANT_NETWORK_TYPES=vxlan,gre,vlan not setting up vlan. The CONFIG_NEUTRON_ML2_TENANT_NETWORK_TYPES is comma separated ordered list of network_types to allocate as tenant networks. And if there was more than one option in list packstack previously didn't set up anything than the first value.
Verified NVR: openstack-packstack-2014.2-0.13.dev1395.gaabe0a2.el7ost.noarch Following to comment #15, Tested OK. Following to that Bug I filed Bug #1181131
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHEA-2015-0155.html