yahoo-eng-team team mailing list archive

[Zhipeng Shen <1671379@xxxxxxxxxxxxxxxxxx>]

** Also affects: neutron
   Importance: Undecided
       Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1671379

Title:
  The first VM of one network in one compute node cannot send RARP
  packets during KVM's live-migration in a neutron ML2 hierachical port
  binding environment whose second mechanism driver was configured as
  the existing OVS driver  "openvswitch"

Status in neutron:
  New
Status in OpenStack Compute (nova):
  New

Bug description:
  Description
  ===========
  Normaly, VM which migrates to destination node can send several RARP packets during KVM's live-migration in my openstack environment.

  In neutron ML2 hierarchical port binding environment,
  I find that the physical port associated to a vlan physical provider's ovs bridge on destination node cannot dump any rarp packets when VM migrates to destination node.

  Steps to reproduce
  ==================
  1. create a vxlan type network:   netA
  2. create a subnet for netA:      subA
  3. create a vm in compute1 node:  vmA
  4. tcpdump the physical port associated to a ovs bridge in compute2 node:
    tcpdump -i ens33 -w ens33.pcap
  5. live migrate the vm to the other compute node: compute2 node
  6. open ens33.pcap in wireshark

  Expected result
  ===============
  find several rarp packets

  Actual result
  =============
  find not any rarp packets

  Environment
  ===========
  OpenStack：Kilo  2015.1.2
  OS: CentOS 7.1.1503
  Libvirt：1.2.17

  Logs & Configs
  ==============
  hierarchical port binding configuration:
  controller node:
  #neutron
  /etc/neutron/plugins/ml2/ml2_conf.ini
  [ml2]
  type_drivers = vxlan,vlan
  tenant_network_types = vxlan,vlan
  mechanism_drivers=ml2_h3c,openvswitch
  #ml2_h3c, a mechanism driver owned by New H3C Group which is a provider of New IT
  #solutions , allocates dynamic vlan segment for the existing mechanism driver
  #"openvswitch"

  [ml2_type_vlan]
  network_vlan_ranges = compute1_physicnet1:100:1000, compute2_physicnet1:100:1000
  [ml2_type_vxlan]
  vni_ranges=1:500

  compute1 node:
  #neutron
  /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini
  [ovs]
  bridge_mappings=compute1_physicnet1:br-ens33

  compute2 node:
  #neutron
  /etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini
  [ovs]
  bridge_mappings=compute2_physicnet1:br-ens33

  Analysis
  ==============
  After reading the live-migration relevant code of nova, neutron-server and neutron-openvswitch-agent, I think that it may be a bug.

  The brief relevant process：

  1. source compute node(nova-compute)  compute1 node
    self.driver(libvirt).live_migration
     dom.migrateToURI2 ---------------Excecute migration to dest node
      self._live_migration_monitor------------------ Monitor migration finished
       self._post_live_migration ---------------- Migration finished
        self.compute_rpcapi.post_live_migration_at_destination  --- Notify
                                                             destination node

  2.1. destination compute node (neutron-openvswitch-agent)   compute2 node
     rpc_loop   ------ monitor vm's tapxxxx port plug
      self.process_network_ports
       self.treat_devices_added_or_updated
        self.plugin_rpc.get_devices_details_list  -------The port details shows that
                   the port still is bound to
                   "compute1_physicnet1", not the physical network
                   provider "compute2_physicnet1" existing in
                    destination compute node.
          self.treat_vif_port
           self.port_bound
            self.provision_local_vlan ---  There is not matched physical bridge at
                   the time. As a result, the tap port can not been set any
                   vlan tag.Eventually, br-ens33, the physical bridge,
                   drops rarp packets from the starting vm.

  2.2 destination compute node (nova-compute)   compute2 node
     post_live_migration_at_destination   nova/compute/manager.py
      self.network_api.migrate_instance_finish
       self._update_port_binding_for_instance ------------Notify neutron migrate port
                                 binding:host_id

  3. controller node(neutron-server)
     ml2_h3c: fill self._new_bound_segment and self._next_segments_to_bind with
          compute2_physicnet1 for openvswitch driver
     openvswitch: bind port with compute2_physicnet1's allocated segment from level 0
            driver ml2_h3c

  In the current process of kilo, ml2 driver finishes port bind at the last step 3.
  it's too late to make neutron-openvswitch-agent get suitable port details from
  neutron-server to set correct vlan tag for vm port and adds relevant flow for ovs bridges
  that nova  notifies neutron-server the event that port changes binding_hostid in ml2
  hierarchical port binding.

  It seems that liberty, mitaka exists the same problem.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1671379/+subscriptions