Failover Cluster Network Design with Hyper-V–How many NICs are required?

Failover Clustering reliability and stability is also “strongly” dependent on the underlying networking design and *drivers* but that’s a another story…. Let’s focus here in the design part.

Since there is no more the hard requirement like in Windows 2003 based clusters (MSCS) for an “HB” (heartbeat) network there is some “unsureness” around the network design for Failover Clustering based on Windows 2008 especially when virtualization workloads are involvedSmileHeartbeat traffic.

Cluster intra-communication (heartbeat traffic) will now go over each cluster network per default except you disable for cluster usage like in case of ISCSI:

image

NOTE: It is a well known best practice to disable cluster communication for ISCSI networks = dedicated for ISCSI traffic only!

The “golden” rule here is, for a “general” Failover Cluster, Microsoft does “recommend” to have at minimum 2 redundant network “paths” between the cluster nodes . But often you want to use more than the minimum “recommended” as you want to have additional redundancy (and/or performance) in your network connectivity (a.e. NIC Teaming) or you will use features like Hyper-V (CSV, LM) which will bring his own network requirements.

Depending on the used workloads on top of Failover Clustering, the number of required physical NICs can grow fast. In example in Hyper-V Failover Clustering with using Live Migration and ISCSI for VM guests the recommended number is roughly at minimum 4 physical NICs, of course more are required, when using NIC teaming technologies for redundancy and or performance objective.

Here are a few example scenarios and the number of the “minimum recommended” required physical NIC ports per cluster node:

Scenario 1:

Failover Cluster with 2 Nodes and Hyper-V (1 x  Virtual Switch dedicated) in use without LM/CSV

image

=> min. 3 physical NICs are recommended => 2 Cluster Networks are automatically discovered and added to Cluster

Scenario 2:

Failover Cluster with 2 Nodes and Hyper-V (1 x Virtual Switch dedicated) in use with LM/CSV

image

=> min. 4 physical NICs are recommended => 3 Cluster Networks are automatically discovered and added to Cluster

Scenario 3:

Failover Cluster with 2 Nodes and Hyper-V (1 x Virtual Switch dedicated) in use with LM/CSV and ISCSI at host

image

=> min. 5 physical NICs are recommended (see note below for ISCSI) => 4 Cluster Networks are automatically discovered and added to Cluster

Scenario 4:

Failover Cluster with 2 Nodes and Hyper-V (2 x Virtual Switch dedicated) in use with LM/CSV and ISCSI at host and guest

image

=> min. 6 physical NICs are recommended (see note for ISCSI) => 4 Cluster Networks are automatically discovered and added to Cluster

Scenario 5:

Failover Cluster with 2 Nodes and Hyper-V (3 x Virtual Switch dedicated) in use with LM/CSV and ISCSI at host and guest

image

=> min. 7 physical NICs are recommended (see note for ISCSI) => 4 Cluster Networks are automatically discovered and added to Cluster

NOTE: In case of ISCSI it is recommended to have at minimum 2 physical network paths for redundancy (availability) purposes. NIC TEAMING IS NOT SUPPORTED HERE, MPIO or MCS must be used for reliability and availability purposes. As a best practice you should disable “cluster communication” through the ISCSI interfaces!

Of course, now when you use techniques like NIC teaming for networks like “Management, Hyper-V switches, CSV..” the number of required physical NICs will automatically grows.

Generally, the cluster service – “NETFT” network fault tolerant – will automatically discover each network based on their subnet and add it to the cluster as a cluster network. ISCSI networks should be generally disabled for Cluster usage (cluster communication).

Further official guidance around network design in Failover Clustering environments can be found here:

Network in a Failover Cluster
http://technet.microsoft.com/en-us/library/cc773427(WS.10).aspx

Network adapter teaming and server clustering
http://support.microsoft.com/kb/254101

Hyper-V: Live Migration Network Configuration Guide
http://technet.microsoft.com/en-us/library/ff428137(WS.10).aspx

Requirements for Using Cluster Shared Volumes in a Failover Cluster in Windows Server 2008 R2
http://technet.microsoft.com/en-us/library/ff182358(WS.10).aspx

Designating a Preferred Network for Cluster Shared Volumes Communication
http://technet.microsoft.com/en-us/library/ff182335(WS.10).aspx

Appendix A: Failover Cluster Requirements
http://technet.microsoft.com/en-us/library/dd197454(WS.10).aspx

Cluster Network Connectivity Events
http://technet.microsoft.com/en-us/library/dd337811(WS.10).aspx

Understanding Networking with Hyper-V
http://www.microsoft.com/download/en/details.aspx?amp;displaylang=en&displaylang=en&id=9843

Achieving High Availability for Hyper-V
http://technet.microsoft.com/en-us/magazine/2008.10.higha.aspx

Windows Server 2008 Failover Clusters: Networking (Part 1-4)
http://blogs.technet.com/askcore/archive/2010/02/12/windows-server-2008-failover-clusters-networking-part-1.aspx
http://blogs.technet.com/askcore/archive/2010/02/22/windows-server-2008-failover-clusters-networking-part-2.aspx
http://blogs.technet.com/askcore/archive/2010/02/25/windows-server-2008-failover-clusters-networking-part-3.aspx
http://blogs.technet.com/askcore/archive/2010/04/15/windows-server-2008-failover-clusters-networking-part-4.aspx 

Description of what to consider when you deploy Windows Server 2008 failover cluster nodes on different, routed subnets
http://support.microsoft.com/kb/947048

Stay tuned…. Winking smile

Regards

Ramazan

Advertisements

Tags: , , ,

8 Responses to “Failover Cluster Network Design with Hyper-V–How many NICs are required?”

  1. George Says:

    Good article for hyper-v cluster.

    one question,
    something not very clear for me.

    Scenario 1: one NIC for hyper-v (no ip)
    what doses mean? one NIC in hyper-v server without ip?
    but what’s purpose for this NIC?

    In Scenario 4:
    v swith1 no IP, so what’s purpose for this nic in child zone?

    Scenario 5:
    one Vswitch scsi without IP, why we need this?

    THX,

    George

    • Ramazan Can [MVP Cluster] Says:

      Hi George,

      here we go with your questions:

      Scenario 1: one NIC for hyper-v (no ip)
      what doses mean? one NIC in hyper-v server without ip?
      but what’s purpose for this NIC?

      Ramazan: “one NIC for hyper-v (no IP)” means you do not share the Network Adapter with host which is the Checkbox I described in my article. As a best practice you should always dedicate the adapter to Hyper-V, where all procotcols (incl TCP/IP) will be removed and only the “virtual Switch protocol” will be bound to this Adapter.

      In Scenario 4:
      v swith1 no IP, so what’s purpose for this nic in child zone?

      Ramazan: as in Scenario 1, this is an dedicated adapter for Hyper-V where no IP is required on host side, as all protocols like TCP/IP will be unbound.

      Scenario 5:
      one Vswitch scsi without IP, why we need this?

      Ramazan: same here, in this Scenario you do provide ISCSI network connectivity to the virtual machines and dedicate the Adapter for hyper-v Networking only, then again no IP on host is required.

      Hope that makes sense, if not please let me know 😉

      Regards
      Ramazan

  2. Mahesh Says:

    Hello Sir,
    I am unable to understand how to assign adapters for CSV and Live Migration.Where did I can find these settings in Hyper-V console ?

  3. Denis Anselmo Ozorio Says:

    Hi,

    I have 2 VM`S in my cluster in a DMZ, but in a failover case, when this VM move to another note, they use a wrong NIC, i need set walys use my DMZ NIC.

    DO have any ideia ?

    Tks
    Denis

    • Ramazan Can [MVP Cluster] Says:

      Do you mean the wrong virtual NIC or wrong virtual Switch? If Switch please update your VM configuration with a.e. powershell:
      Import-Module FailoverClusters
      Update-ClusterVirtualMachineConfiguration –name “VMNAME”

      If NIC, I do assume you have an R2 SP1 cluster and fully patched?

  4. Bryan Says:

    Great Article

  5. Failover Cluster Network Design with Hyper-V – How many NICs are required? | /dev/random Says:

    […] Here are a few example scenarios and the number of the ?minimum recommended? required physical NIC ports per cluster node: source […]

Any further thoughts? Let me know here

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: