Want to protect your cyber security and still get fast solutions? Ask a secure question today.Go Premium

x
  • Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 83
  • Last Modified:

Hyper V Failover Cluster 2012 R2 Cluster Teaming Question

Hi All,

I am currently in the process of building a new hyper v cluster so i can migrate from 2008R2 to 2012R2. On my 2008R2 cluster we have 1 single nic for live migration and 1 nic for the privacy cluster communication, i know i can now team the live migration, i would like to also team the privacy cluster nic so it has hardware redundancy across 2 nics. Would this have any negative impact on the cluster communication?

Thanks in advance

Danny
0
WycombeAbbey
Asked:
WycombeAbbey
  • 3
  • 2
1 Solution
 
Philip ElderTechnical Architect - HA/Compute/StorageCommented:
How many NIC ports per node and what manufacturer?
0
 
WycombeAbbeyAuthor Commented:
Ok So....

We have 16 NIC ports 4 of which are Broadcom BCM57810 Netextreme II 10 Gig E and 12 Broadcom Netextreme Gigabit Ethernet.

I plan to use them as below:

2 x Gigabit Broadcom for Management connection to the host server (Teamed)

4 x Gigabit Broadcom for VM traffic (Teamed)


2 x Gigabit Broadcom for Live Migration (Teamed)

2 x Gigabit Broadcom for Privacy Cluster Comms (Teamed if Possible)

2 x 10 Gig Broadcom for isci traffic, 1 of each going to seperate isci switch for redundancy

I hope this helps.

Thanks

Daniel
0
 
Philip ElderTechnical Architect - HA/Compute/StorageCommented:
Wow ... our default is 8 Gigabit and 4 10GbE all on Intel (i350-T4 x2 and x540-T2 x 2).

We would do the following:
Team 0: Port 0 on NIC 0 + 1 + 2 = Management OS
Team 1: Port 1 + 2 on NIC 0 + 1 + 2 = EXCLUSIVE vSwitch (Virtual Switch not shared with host OS)

That's 3 ports for the management OS + 6 ports for the vSwitch leaving 3 ports.

No Team:
Port 3 on NIC 1: Live Migration Dedicated (192.168.100.0/24)
Port 4 on NIC 2: Live Migration Dedicated (192.168.110.0/24)

The extra port could be set up for Live Migration or left to cluster communications.

The 10GbE ports should be dedicated to iSCSI traffic only with Jumbo Frames enabled throughout the network stack and in the switches.

I recommend a pair of Gigabit switches to provide some redundancy on that side of things too.

Also, for the Broadcom ports the vSwitch gets bound to make sure to disable Virtual Machine Queues _in the driver_ not the OS as network performance will suffer big time otherwise.
0
 
WycombeAbbeyAuthor Commented:
Ok Thanks for that, so you would recommend teaming cluster comms or having more than 1 nic across more than 1 card, dedicated to the cluster comms as in the example of live migration?
0
 
Philip ElderTechnical Architect - HA/Compute/StorageCommented:
If the production network team have two separate switches, that is one per port, then leaving cluster comms enabled there means they are redundant.

Other than two paths to keep things resilient to outages we don't configure dedicated networks beyond the production, vSwitch(es), and Live Migration networks. There really is no longer a need to in the 2012 R2/2016 era.
0

Featured Post

What does it mean to be "Always On"?

Is your cloud always on? With an Always On cloud you won't have to worry about downtime for maintenance or software application code updates, ensuring that your bottom line isn't affected.

  • 3
  • 2
Tackle projects and never again get stuck behind a technical roadblock.
Join Now