none
Failovercluster Management Network Interface RRS feed

  • Question

  • Hello Friends,

    my question today is revolving around Failoverclustering (which is sitting, in our case, on top of a S2D deployment). Everything is working fine in that regard.

    Next week we will get new switches and I plan to trunk/lacp both SFP+ ports on the servers (at the moment only one SFP+ port of every server is connected to the switch) and connect them to the new switches.

    Iam aware how trunking/lacp works on the powershell, thats not the problem. What I fear is the connectivity of the failovercluster itself. When I have to dissolve the Hyper-V Virtual Ethernet Adapter to form a new LACP, so I can create a new Virtual Adapter on top of the LACP, the failovercluster will lose the only management interface. 

    When I look into the Failovercluster MMC Snap-In under "Network" I can see my interfaces, but I dont see any buttons to add interfaces and declare them as management network (my idea was to use the copper ports of the server nic to create a temporary management interface until I have built the LACP/virtual adapter and integrate them as the primary management interface).

    So the question remains: can I add (maybe through powershell) additional interfaces as management to the failovercluster or do I actually have to dissolve the whole cluster and build anew?

    Thanks in advance.

    Best Regards,

    Constantin


    • Edited by synth3tix Thursday, August 22, 2019 8:06 AM
    Thursday, August 22, 2019 8:05 AM

Answers

  • You can add new interfaces to a cluster - no problem.  Each host in a cluster recognizes all the NICs that it owns.  It is a matter of using management tools to define them for use to the cluster (or remove them from use).

    tim

    • Marked as answer by synth3tix Friday, August 23, 2019 10:49 AM
    Thursday, August 22, 2019 12:56 PM
  • Hi,

    Thanks for your question.

    I agree with Tim, you can add new interfaces to a cluster. It is one of great improvements in Windows 2008 failover clustering that the cluster automatically detects the available networks (NETFT) and add them to the cluster. Just install/enable them + assign IP and wait some minutes till the the networks are automatically added. BUT you should run anyway an "network only" validation test to verify your newly added IPs and their network configuration on each node. (http://technet.microsoft.com/en-us/magazine/2008.07.failover.aspx ) The only important side note here is, that each cluster network MUST be in seperate VLAN/subnet/network.

    Therefore, it wouldn't impact your cluster. But it's always good to do a validation after adding the NICs. We can run validation and select only the Network tests.

    Hope this helps.

    Highly appreciate your effort and time. If you have any question or concern, please feel free to let me know.

    Best regards,

    Michael


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

    • Marked as answer by synth3tix Friday, August 23, 2019 10:49 AM
    Friday, August 23, 2019 7:31 AM
    Moderator
  • "the other is for the Heartbeat"

    Just an FYI … there is no separate 'heartbeat' network.  Heartbeat is just one type of packet that travels on the cluster communication network.  Any network defined as available to carry cluster communication will carry heartbeat traffic.  It is always a good idea to have at least two NIC defined to carry cluster communication or to define that network on a teamed interface.


    tim

    • Marked as answer by synth3tix Monday, August 26, 2019 7:38 AM
    Friday, August 23, 2019 1:40 PM

All replies

  • You can add new interfaces to a cluster - no problem.  Each host in a cluster recognizes all the NICs that it owns.  It is a matter of using management tools to define them for use to the cluster (or remove them from use).

    tim

    • Marked as answer by synth3tix Friday, August 23, 2019 10:49 AM
    Thursday, August 22, 2019 12:56 PM
  • Hi,

    Thanks for your question.

    I agree with Tim, you can add new interfaces to a cluster. It is one of great improvements in Windows 2008 failover clustering that the cluster automatically detects the available networks (NETFT) and add them to the cluster. Just install/enable them + assign IP and wait some minutes till the the networks are automatically added. BUT you should run anyway an "network only" validation test to verify your newly added IPs and their network configuration on each node. (http://technet.microsoft.com/en-us/magazine/2008.07.failover.aspx ) The only important side note here is, that each cluster network MUST be in seperate VLAN/subnet/network.

    Therefore, it wouldn't impact your cluster. But it's always good to do a validation after adding the NICs. We can run validation and select only the Network tests.

    Hope this helps.

    Highly appreciate your effort and time. If you have any question or concern, please feel free to let me know.

    Best regards,

    Michael


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

    • Marked as answer by synth3tix Friday, August 23, 2019 10:49 AM
    Friday, August 23, 2019 7:31 AM
    Moderator
  • Hello Tim, Hello Michael,

    thank you both very much for your replies.

    Now that you guys phrase it like that, I remember something. When I first deployed the S2D/Failovercluster, there was no "adding". I defined the interfaces and then they appeared in the Snap-In. 

    Yeah.

    Since I want to reforge the current network segmentation anyway, I'll make full use of the copper ports (well, one port of the Intel X710, the other is for the Heartbeat) and will refine my network structure in that regard.

    Thanks again!

    Best Regards,
    Constantin

    Friday, August 23, 2019 10:49 AM
  • "the other is for the Heartbeat"

    Just an FYI … there is no separate 'heartbeat' network.  Heartbeat is just one type of packet that travels on the cluster communication network.  Any network defined as available to carry cluster communication will carry heartbeat traffic.  It is always a good idea to have at least two NIC defined to carry cluster communication or to define that network on a teamed interface.


    tim

    • Marked as answer by synth3tix Monday, August 26, 2019 7:38 AM
    Friday, August 23, 2019 1:40 PM
  • Hey Tim,

    thank you for the information. Will keep that in mind and reorganize the cluster interfaces when these new switches arrive.

    Best Regards,

    Constantin

    Monday, August 26, 2019 7:45 AM