vSphere Replication

 View Only
  • 1.  Doubts over isolating VR traffic

    Broadcom Employee
    Posted Feb 01, 2023 01:35 AM

    Hi all,

    I am trying to isolate the VR traffic in my lab but I have 3 doubts over the setup:

    1) In the "IP address for incoming storage traffic" is it correct that you need to enter the IP address of the new network adaptor that you added for separating the VR traffic.

    e.g. if you create a new VMkernel adaptor for VR traffic with an IP of 10.10.10.10 and connect that port group to the VR appliance, that is the IP you would put in that field.

    I believe that you are basically telling the VR appliance to accept all VR traffic through that VMkernel instead of the mgmt kernel, correct?

     

    2) When adding a static route on the source VR appliances, you enter the default gateway that the network needs to use to go to the target site. Is that correct?

    Also you enter the target network IP address subnet as well.

     

    3) In the documentation it says in one place that you should only put 1 service on a single VMkernel adaptor but in another place that you can enable both services (VR and VR NFC) on 1 adaptor.

    Can you enable both services on a single adaptor without having any problems or being unsupported?

     

    Can anyone confirm my 3 doubts please?

    Cheers



  • 2.  RE: Doubts over isolating VR traffic

    Broadcom Employee
    Posted Feb 03, 2023 07:53 PM

    I think I have found the answers.

    1) "In the IP address for incoming storage traffic", you enter the IP address of the VMKernel adaptor that you have configured for vSphere Replication traffic on that appliance

    2) "When adding a static route on the source VR appliances, you enter the default gateway that the network needs to use to go to the target site":

    On the VR appliance you enter the following section in the VR traffic eth1 adaptor file (10-eth1.network):

    [Route]

    Gateway=<gateway for the VR traffic subnet>

    Destination=<ip subnet range of vr-traffic subnet at DR site> in CIDR format e.g. 192.168.0.1/24

    FYI it seems to be better to fully separate the traffic rather than just the VR traffic so you would need to add 2 NICs to the VR appliance (one for VR traffic - inbound traffic - and one for VR NFC traffic - outbound traffic).

    If you add 2 NICs then the network directory on the VR appliance will have 3 files like this:

    10-eth0.network     for the mgmt traffic

    10-eth1.network     for the VR inbound traffic

    10-eth2network     for the VR NFC outbound traffic.

    3) As already mentioned it seems to be more recommended to separate all traffic so you would use 2 port groups with 2 VMKernel adaptors instead of 1 VMkernel adaptor with both types of traffic enabled.

    Also don´t forget that you actually need to add static routes (if required) to both the ESXi hosts AND the VR appliances.  -> a static route in 2 places.

     

    I hope this info helps someone.

    Cheers