Article written by

Sylvain is a principal software engineer at Red Hat. His main interests are OpenStack Neutron and SDN solutions. At his spare time he likes to play tennis and badminton.

8 Responses

  1. OpenStack Community Weekly Newsletter (July 17 – 24) - GREENSTACK

    […] Sylvain Afchain: A journey of a packet within […]

  2. Klaus Samardžić
    Klaus Samardžić at |

    Nice article. It is always helpful to have a visual (and Wireshark) presentation of what is happening in the network.


  3. yatin
    yatin at |

    For MPLS Label 16 the agent introspect,
    http://:8085/Snh_MplsReq, shows nh_index: 13


    $mpls –get 16 shows nexthop 14

    I’m confused, to me it looks mismatch.

    but, correct me in case wrongly interpreted?

  4. yatin
    yatin at |

    Great Article. Kudos to you 🙂

    even if we have two VMs on two compute nodes, from same subnet.

    Vrf id on both compute node differs.
    like on compute-1 Vrf:5 and compute-2 Vrf:13

    there’s no one-to-one mapping between virtual-network vrf-id

    in another words, if VMs from same virtual-network, span across compute nodes, all compute nodes will have different vrf-id to refer to same virtual-network.

    how does that managed by control node?

    Also, I haven’t come across, detail (ex: data-flow / packet-flow /or how to debug connectivity) technical article on control-control, vrouter-agent nd gateway routers.

    May be you can, write another blog post towards this, that will be equally helpful as this one.

  5. yatin
    yatin at |

    Thank You Sylvain.

    eagerly waiting for your next post. 🙂

  6. Ravi Prasad
    Ravi Prasad at |


    I am not able to establish ping among VMs in case of following topology :

    LAN-A Subnet –
    LAN-B Subnet –
    LAN-C Subnet –

    I am trying to ping VMA from VMC. I am able to ping VRouter2 interface ( from VMC . But VMA is not reachable from VMC. Please let me know how to make this work. Do I need to do some additional configuration(like static route etc.) ?

    I used “vif”/”rt”/”nh” and tcpdump on eth0 as mentioned in above blog. I observed that packet is not coming out of eth0. However tcpdump on tap interface confirms that VM has sent out pkt to host. Please let me know if we need to add default route in this case. If yes, how we can add default route in vRouter?

Comments are closed.