Hyper-V NIC Teaming & Virtual Switches

in this video we're going to be talking about NIC teaming and virtual switches as they pertain to hyper-v networking we'll be talking you through how you can make the best of the available NICs that you have whilst increasing bandwidth in the traditional hyper-v module what'll it be assigning different NICs to different traffic types such as management cluster harpie CSV etc and we do this to guarantee that bandwidth was available to each of the networks when it needed it so we'd have a couple of NIC society management I say a couple because we'd want to team them together to make sure that the network was highly available we separate out the plus heartbeat CSV like migration we'd promote networking face cards and some bandwidth for users to collect into virtual machines and we'd even have network cards not necessarily team together but may be presented by MPI o4i scuzzy or connecting to a fiber Channel SAN in order for the virtual machines to sit hardest to sit on external storage as you can see this takes quite a few Nicks quite a few ports in a data center quite a bit a configuration pipe air management the choice that you do have in more modern world is to use conversion networking where we use fewer Knicks but split up the bandwidth available on those next to the host operating system into the virtual machines for example you may be using physical 10 gig cards and those cards may allow us to split network up at the physical card level and assign quality of service if you bought the wrong cards you may not get to do this is you mean you have 10 gig cards or indeed if you just have a one gig network then you face a choice of either using lots of lots of ports or you can bond those cards together into a large team with more than two NICs and then split up that pool of bandwidth between the network such as management that's the heartbeat CSV and live vibration aho with Windows Server 2012 r2 you can also assign Paulito service policies to those networks to guarantee them either a minimum or a maximum amount of bandwidth and control the data flow through the networks so let's go and have a look in the lab as you can see in the lab I have three network cards configured one for the LAN one for CSV and cluster one for live migration there's an extra V Ethernet card and that's there because my hyper-v switch major is shared with my management layer over in further of cluster manager you can see that we have three networks configured one for each network we only have LAN live migration and CSV cluster matching the three network cards that we have and the three subnets that we're using 192 168 1 2 & 3 so what I'd like to do is recover these network cards and team them together and then present those up to a virtual switch and from there create virtual networks for our management operating system to use so the first thing that we'll do is we'll remove any network information from these network interfaces so we'll take that out and from the live migration we'll remove the information there obviously this needs a bit of planning if you're going to do doing this in production well then rename these to team NIC 1 or whatever naming convention you would like to use in your environment for identifying the different NIC team members let's rename that to team NIC - okay so now I have to nix free to create as a team if we access server manager see the nic team is currently disabled and there are no teams there i can select to nix and from there we could add to a new team and if we look at the additional properties we'll see that we have some selections that we can make a teaming mode can be switch static teaming switching an or lack P load balancing mode use address hash hyper-v port or dynamic and we can select use a standby adapter in the lab we have our three network cards as we say then team Nick 1t nip - and I'm going to create a team and I could come in select the two Nick's that I want to be in my team and then create a new team but this is the brave new world where use PowerShell to automate tasks and the reason we use PowerShell quite frankly as if you're building more than one server then errors can creep in a few performing tasks manually so it's quicker and more certain if you use PowerShell that you'll carry out the same steps exactly on each server so hopping into PowerShell we can arm typing get adapter and that will list our adapters and we should see I have renamed adapters Teaneck one Technic to seeing there so I could have used if I had liked arm rename net adapter and one call let's say for example team Nick - and I could have team neck - and we could have named that my new name and if we press turn go back and refresh our interface then we'll see that Nick is instantly renamed so let's turn those around my new name better teennick - my new name rename that back to Teen Nick - come back refresh the screen and hey presto change-o and nixon named as they should be so let's create that new load-balancing failover team so new net load balancing failover team and we'll give it a day and the name that we will give this is management team and age month t and the team members if I press tab it autocomplete the entries for me so that will be team Nick one and t Nick to the teaming mode then I'll put them into is switch independent and the load balancing algorithm if you remember below bouncing algorithm choices that we had or address hash hyper-v mode or dynamic so I'm going to follow the current press practice the new best practice and load balance them dynamically and because I don't want the message saying do you really want to do this I will put confirm force and that will create a new team and in the background you see the networks we connect three converging and the management team created and if we wait there then that Nick should eventually reconverge and appear for us but while waiting for that we'll move on so we have a new network created a new team created but in our hyper-v we certainly have the one switch so what we do next after grating the team is we'll create a new virtual machine switch which will call our converged network switch and a minimum bandwidth mode will be based on the weight and then net adapter name that we will connect that to is our management team the network card that we just created the network team Carter we just created and I don't want to connect that up to our management interface so I press return now that should go away and create a new switch for me if I go back into my switch manager there is my new switch and note that we not connected to the management network and once I entered in the name of the network I wanted to connect to the management team name of Nick it's connected me up to the correct one of my drum one of my correct one of my cards so at the moment we have a team and we have a virtual switch but what we don't have already additional networks so let's create those the command went here would be to add a new VM network adapter and I'm going to add that in the management OS and I'm going to call that live migration and I'm going to hook that up to my switch named converged converge their switch and the media in the backgrounds you see light migration being created I'm then going to add a new network adapter and this one I'm going to name cluster CSV so slightly different from our previous one which had a space in the - but that's because makes coding so much easier fact what I will do is I will have a cluster network and there's my one for cluster communications and then I can have a CSV Network and so what you can see whereas before we only had two networks running over these two network cards we now have three networks running over the two network hearts such created an interface for us and on those you'll notice that the settings and properties are very much the same as for a standard card and we could now go back in that in our addresses so let's quickly do that and say that our live migration we'll give back it switching the dress one nine two one six eight dot and cluster communications sorry as CSV will give the original address that that had which is 192.168.0 and then what we'll do is we'd create a new VLAN on our switch on line 2 1 6 8 4 2 4 take those out and then we have 3 networks but what they'll all do is try to use all of our bandwidth so what we'll do now so we have quality service to each of those networks that we've created so if we set a VM network adapter in the management OS the one named live migration and minimum band width by weight so we have absolute we have weight absolutely set absolute minimum bandwidth weight will give a waiting value if those values all add up to 100 then you can think of it as being a minimum percentage of the available bandwidth so let's free live migration give it a value of I don't know if 30 and if more than 3% is available then Network will use it but if that's all that's available then that's all they'll use so we'll give 20% to our cluster communications and then if our CSV need to be running its VMS across the network or controlling the CSVs across the network then we'll give that half of the bandwidth that we have available to us and say if more than that is available then it will use more and now we've applied course to our settings so if you remember at the start of that time in the lab we only had these three NICs available to us could only do three networks and we still have the three NICs available to us but we now have an additional network running through this we have four networks back in that converge switch see that we are sharing that networks with a management the switch with the management layer but as you can see that's all grayed out we can't change it because that's being done a very much lower level I hope you enjoyed watching this and look forward to seeing you in their next video

Channel: SystemCentric

Views: 57,723

Keywords: Hyper-V

Id: SG3CfX52OSI

Length: 16min 5sec (965 seconds)

Published: Sun Jul 13 2014