Wired Intelligent Edge

 View Only
  • 1.  VRRP

    Posted Jun 23, 2021 11:18 AM
    I have built a small lab to study VRRP. I built it in EVE-NG Professional
    I have four ArubaOS CX Virtual.10.06.0110 running

    Two are working as layer 3 switches
    Two are working as layer 2 switches

    I have two Docker eve-gui-servers running linux

    I have two vlans configured... vlan 20 and vlan 30

    I setup the vlans, vrf tables, and VRRP

    I am able to ping the DG on both vlans, and the two eve-gui-servers can ping each other. Everything is working, but REALLY slow. 

    I did a wireshark capture from the eve-gui-server ethernet interface on vlan 20 and I am seeing ~850 multicast VRRP packets a second.  I think there should be one per second. I am thinking I have a setting wrong that should be 1 sec, or 1000 ms or something. And I am only seeing about 850 a second because there is not enough CPU on my EVE box to send 1000 a second out. Am I close? I cant find the setting for this, anyone have any ideas? And assistance would be great. Thanks.

    ------------------------------
    Daniel Worden
    ------------------------------


  • 2.  RE: VRRP
    Best Answer

    Posted Jun 24, 2021 09:52 AM
    Is your vrrp in a classic triangle with STP. If so a known issue exits when passing layer 3 traffic over STP loop setup  on the OVA. This gives some loss or intermittent communication.

    ------------------------------
    Kamal Takodra
    If my post was useful accept solution and/or give kudos
    ------------------------------



  • 3.  RE: VRRP

    Posted Jun 24, 2021 10:07 AM

    My config looks like this...

     

    So I think that might be what is happening. Should I block multicast on some of the links like 1/1/2 on all 4 switches? I think I recall seeing something about multi-cast and STP/MSTP. I will look for that article again.

     

    Thanks for your time, it is appreciated.

     

    Dann Worden

     






  • 4.  RE: VRRP

    Posted Jun 24, 2021 06:35 PM

    @ktakodra,

    I ran a test and deleted all the connections between the switches that uses 1/1/2, removing all the loops and redundancy.

    After doing this I only see ~ multicast VRRP packets per second. So, what you were referring to looks more likely to be the case. I am looking for the article I saw on blocking multicast traffic on links in a MSTP environment. I think then I can add back the links and this issue will be resolved.

     

    On a side note, my CPU time is WAY down without all those multicast packets going round.

     

    Thanks again for this input. Really was helpful and appreciated.

     

    Dann