none
RDMA configuration

    Question

  • Hello

    i have hyper v server and 5 vms on it. Please tell how i can configure RDMA on Hyper v server. thanks

    Friday, March 3, 2017 10:14 PM

Answers

  • Hi Sir,

    Both "vm will restart" and "take some time to restart"  are normal behavior .

    I'd like to call it "unplanned" failover .

    The cause is the memory of VMs running on the failed node will be lost while the cluster node experiencing unexpected power outage .

     

    Best Regards,

    Elton


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.


    Sunday, April 9, 2017 9:45 AM
    Moderator
  • "Any solution for this? i dont want to take down time."

    That is the nature of a failover cluster.  A VM running on nodeA is using physical memory on nodeA.  If nodeA fails unexpectedly, there is absolutely no way for the memory image of the VM running on nodeA to be transferred to another node in the cluster.  Memory transfer requires that nodeA be operational.

    There are a couple solutions that will mirror the memory between two hosts.  This means that a VM running on one node is using equivalent resources on a second node.  If you really need this capability, you will pay for it, as it does not come cheaply.  Stratus and Double-Take are two solutions that I have seen over the years.


    tim

    • Marked as answer by chapter 7 Tuesday, April 11, 2017 8:19 AM
    Monday, April 10, 2017 2:18 PM

All replies

  • Hi Hassan,
    What OS you have in your Hyper-V server?
    In Windows Server 2012 R2, using both RDMA and Hyper-V on the same computer as the network adapters that provide RDMA services can not be bound to a Hyper-V Virtual Switch. This increases the number of physical network adapters that are required to be installed in the Hyper-V host. In Windows Server 2016, you can use fewer network adapters while using RDMA with or without Switch Embedded Teaming (SET).

    Please check the following article:
    https://technet.microsoft.com/en-us/library/mt403349.aspx

    Regards,
    -Tomi

    Tomi Pietilä

    Please remember to mark the replies as answers if they help and unmark them if they provide no help.
    Blog
    Twitter
    LinkedIn

    Saturday, March 4, 2017 4:08 AM
  • i am using windows server 2016 and hyperv is installed on it.

    i have 2 NICs on windows server. 1 is for vswitch and 1 is for windows server 2016.

    now i will use this command

    Enable-NetAdapterRDMA "Eth0-vswitch"

    it will work?

    Sunday, March 5, 2017 10:05 PM
  • Hi,

    Did you get any updates on the configuration?

    Best Regards,

    Leo


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.

    Thursday, March 30, 2017 1:57 AM
    Moderator
  • Hello

    I have configured it.

    Regards

    Friday, March 31, 2017 12:37 PM
  • Do your network adapters support RDMA?  Not all NICs are RDMA compatible. Use the PowerShell command Get-NetAdapterRDMA to list RDMA-capable adapters.

    tim

    Friday, March 31, 2017 1:56 PM
  • Hi,

    Just want to confirm the current situations.

    Please feel free to let us know if you need further assistance.

    Best Regards,

    John


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.

    Friday, April 7, 2017 9:24 AM
  • Hi John,

    Thanks for asking. 

    I was new in hyper v. thats why i need some help. I am learning it myself. Now i am learning the Topic of hyper v failover cluster with ISCSI. if i need help, surely i will asked you. thanks

    Regards

    Friday, April 7, 2017 9:25 PM
  • Hello,

    i have created hyperv clustered with two hyperv servers.

    when my first hyperv server goes to down then all vms which are created on first hyperv server will goes to second hyperv server.

    there are two problems.

    first- vm will be restart. 

    second- it will take to some time to restart.

    can we have any solution, vm should not restart.

    Regards

    Saturday, April 8, 2017 2:27 PM
  • Hi Sir,

    Both "vm will restart" and "take some time to restart"  are normal behavior .

    I'd like to call it "unplanned" failover .

    The cause is the memory of VMs running on the failed node will be lost while the cluster node experiencing unexpected power outage .

     

    Best Regards,

    Elton


    Please remember to mark the replies as answers if they help.
    If you have feedback for TechNet Subscriber Support, contact tnmff@microsoft.com.


    Sunday, April 9, 2017 9:45 AM
    Moderator
  • hello,

    Any solution for this? i dont want to take down time.

    Sunday, April 9, 2017 11:01 AM
  • "Any solution for this? i dont want to take down time."

    That is the nature of a failover cluster.  A VM running on nodeA is using physical memory on nodeA.  If nodeA fails unexpectedly, there is absolutely no way for the memory image of the VM running on nodeA to be transferred to another node in the cluster.  Memory transfer requires that nodeA be operational.

    There are a couple solutions that will mirror the memory between two hosts.  This means that a VM running on one node is using equivalent resources on a second node.  If you really need this capability, you will pay for it, as it does not come cheaply.  Stratus and Double-Take are two solutions that I have seen over the years.


    tim

    • Marked as answer by chapter 7 Tuesday, April 11, 2017 8:19 AM
    Monday, April 10, 2017 2:18 PM