Recommended management and vmotion redundancy?

[w1ll1ng]

Hi,

Is it recommended to have vmotion and management on the same vswitch for redundancy with atleast 2 NIC's, or have seperate vSwitches for each with dual NIC's per switch minimum?

[a_p_]

It actually depends on the size of your environment. In smaller installations, I usually use one vSwitch with 2 uplinks for Management and vMotion. In each of the two port groups I configure the uplinks as active/standby and vice versa, so that during normal operation, each port group uses it's dedicated physical NIC.

André

[w1ll1ng]

Makes sense,

The thing is though, when you have both on the same switch, I would think there can be saturation on the active link in the event of a failure.  If there is much vmotion or drs activity that is used with vmotion, could this afftect HA network heart beat and management traffic in general or vice versa?  I think GB links may be fine, advice?

[a_p_ · Accepted solution]

I don't expect a saturation where no heartbeat traffic will go through anymore with a gigabit uplink.

Btw. what I forgot to mention in my previous post is that I usually work with different VLANs on the port groups and the physical ports configured as trunk ports.

André

[logiboy123]

What you want to do is pretty much standard practise. You need to make sure that you apply load balancing rules on the uplinks so that vMotion does not saturate the link used for Management.

vSwitch0

Management - vmnic0 Active / vmnic5 Standby

vMotion - vminic0 Standby / vmnic5 Active

See this diagram for a view of what I'm talking about:

http://vrif.blogspot.com/2011/10/vmware-vsphere-5-host-network-design-6.html

Regards,

Paul

[w1ll1ng]

I guess this helps as well.  I have used active/standby with iscsi mpio and nic teaming (trunking) for vm networks.  Seems like a bit of both without NIC bindings.  I guess the vlan will go across the surviving link of trunk and decapsulate fine.  I will try this, the dev is small 2 hosts. thx (to Andre)

[w1ll1ng]

Looks very informative Paul , will have a read and let ya know! thx

[w1ll1ng]

Hi,

Thanks for the recommendations and info.  I have 2 x hosts in a cluster. I used the management/vmoiton redundant switch topolgies as mentioned with seperate vlans etc.  Tested redundancy for either port group and it works well for either service when either switchport/vmnic is disabled.  I was monitoring utilization traffic/usage at both the switchport and esxi level of the vmoiton vmnics/switchports.  From inception I observed that vmotion traffic  traffic was on management network ports for each exsi host. By disabling the management network switchports/vmnic during vmotions it then showed vmotion traffic on the dedicated vmotion vmnics..  By rebooting the hosts numerous times I realized that the vmotion utilization was fluctuating on either management or vmotion vmnics from either hosts,sometimes both being the dedicated vmotion nics or managment and vmotion mixed?  Redundancy does work etc, Is this okay?