7 Responses

  1. Peter
    Peter at |

    Nice write up. I have found the same using different gear. You should also set the link failover detection to beacon probing to allow detection of downed links upstream. This is most important for when you lose the last LAG member on one of the switch member NICs as both links are up from the host perspective.

  2. Josh Odgers (VCDX#90)
    Josh Odgers (VCDX#90) at |

    Great post Chris, clear and straight to the point. I totally agree LBT is the way to go, and you have provided a great example of how to utilize LBT in a converged environment. Keep up the great work.

  3. Eric Shanks
    Eric Shanks at |

    Very nice write up Chris. I’ve seen plenty of questions about VC LAGs and how they relate to vsphere switches.

  4. JP Jung
    JP Jung at |

    Thank you for clearing that nagging question out of my mind! Really a great summary.

  5. tbacks13
    tbacks13 at |

    I am in a UCS environment and tried to implement LAGs on my vDS and it didnt go well. The Cisco core saw MAC flaps and traffic was sluggish at best. I had to revert back to using just the individual uplinks to the vDS; I believe it was because both the physical switch as well as the UCS were using Link Aggregation methods. Any thoughts?


Share your point of view!