Switching

 View Only
last person joined: yesterday 

Ask questions and share experiences about EX and QFX portfolios and all switching solutions across your data center, campus, and branch locations.
  • 1.  VMware ESXi vDS LACP to EX-switch - What config is required on the EX?

    Posted 06-03-2022 06:23
    Hi,

    The goal will be to move our VMware infrastructure from NIC-teaming to LACP. The EX/QFX-side will be MCAE.
    The major question when reading up on the VMware side is the important note that both sides needs to use the same load balancing algo.
    Anyone with Juniper EX/QFX experience that can confirm how you have this running in your environment;
    VMware: Config (IP-based  load balancing?)
    EX/QFX: Anything else except default 802.3ad, slow/fast settings? - From what I've read and can see on our devices, LACP on Juniper doesnt support different hashing algorithms.

    Thanks in advance!
    //Rob


  • 2.  RE: VMware ESXi vDS LACP to EX-switch - What config is required on the EX?

     
    Posted 06-04-2022 10:19
    "both sides needs to use the same load balancing algo."

    Why?



  • 3.  RE: VMware ESXi vDS LACP to EX-switch - What config is required on the EX?

    Posted 06-06-2022 14:25
    We have a UCS mini, and each blade is configured like this to a EX4600 VC.  Works great.  And yes, both sides need to be on the same page,  fast/fast or slow/slow.  Otherwise, earth's axis starts to wobble.

    set interfaces aeX aggregated-ether-options minimum-links 1
    set interfaces aeX aggregated-ether-options lacp active
    set interfaces aeX aggregated-ether-options lacp periodic fast
    set interfaces aeX unit 0 family ethernet-switching interface-mode trunk


  • 4.  RE: VMware ESXi vDS LACP to EX-switch - What config is required on the EX?

    Posted 06-07-2022 05:17
    Thanks for confirmation of the setup, looks straight forward as I hoped for.
    I will proceed to plan for the migration. 
    Regarding the other post from @smicker "why", yes that was pretty much my thoughts as well, but VMware pointed that out pretty clear. I cant see any direct cause of the sides is using diff algorithms to distribute the traffic along the lacp members.