Currently I am working on implementing dual-stack (so all servers and computers will run on IPv4 and IPv6 at the same time) in Windows envrironment with Active directory domain controlllers, other member servers (file server, DFS, SharePoint services…), Exchange server 2013, Lync/Skpye for business…
Purpose of this post is to walk you through the obstacles and difficulties while implementing both protocols to work together…
So basicaly we need to know the folowing:
We have one (or more) public IPv4 addresses which we NAT in our private networks where we have our servers…
Our providers gives us some IPv6 prefix for “wan” interface of our router and over that there is a routed prefix which we will use internaly (you need to know there are public – or globaly routed IPv6 addresses inside your network – SO TAKE CARE of your Firewall roules (We will cover that later)).
Simple steps to implement dual stack is to:
a. Get IPv6 from your provider
b. Have a router that understands IPv6 🙂
c. Configure router advertisment on internal network with M (managed (this will force users to use DHCPv6 instead of autoconfiguring IPv6 (SLAAC)) and O (other configuration (this will point clients to DHCPv6 server to get DNS servers (your domain controllers IPv6 addresses)) flag
d. Configure DHCP server on your Windows server with DHCPv6 parameters (prefix, exclusions, DNS servers (called: 00023 DNS Recursive Name Server IPv6 Address)
e. disable DHCP client on servers that use static IPv4/IPv6 addresses (if you do not do that your servers will autoconfigure additional IPv6 addresses as told by RA…) You can use Powershell: Set-NetIPInterface –InterfaceIndex <number> -Dhcp Disabled
If you have two or more NICs joined in a Team by using Windows server teaming solution and then you use this Team as a base for Virtual switch in Hyper-V and you enable “Allow management operating system to share this network adapter” like:
you will find warnings in system Event log:
A port on the virtual switch has the same MAC as one of the underlying team members on Team Nic Microsoft Network Adapter Multiplexor Driver
Check your NICs by using Powershell cmdlet: get-netadapter | ft Name,MacAddress
You need to change MAC addresses of your Virtual Switch management interface (vEthernet (xxx)) by using Powershell cmdlet: set-netadapter -name “vEthernet (XXX)” -macaddress xx:xx:xx:xx:xx:xx
and MAC addresses of your network cards (in my case 4 NICs) you can leave Team interface mac address alone…
There will be no errors in Event log any more 🙂
I have two HP DL 380 G8 servers with Windows server 2012 R2 OS. I have formed NIC teams by using Windows – built-in NIC teaming.
It Works perfectly but after a week or two teams stop working – the only mode to get server online again is to disable and re-enable physical network cards.
Server uses: HP Ethernet 1Gb 4-port 331FLR Adapter – I have upgraded firmware to latest version (that was available on 1.2.2015) and also updated drivers – but the problem persist.
For a current workaround I have scripted a powershell script that checks connectivity and cycle network adapters:
if (Test-Connection 18.104.22.168 -Count 1 -ErrorAction SilentlyContinue)
Add-Content C:watchdogresult.txt “`nUP”
Disable-NetAdapter -Name “Ethernet” -Confirm:$false
Enable-NetAdapter -Name “Ethernet”
Disable-NetAdapter -Name “Ethernet 2” -Confirm:$false
Enable-NetAdapter -Name “Ethernet 2”
Disable-NetAdapter -Name “Ethernet 3” -Confirm:$false
Enable-NetAdapter -Name “Ethernet 3”
Disable-NetAdapter -Name “Ethernet 4” -Confirm:$false
Enable-NetAdapter -Name “Ethernet 4”
$datenow = Get-Date
$datesult = “`n” + $datenow + ” Repaired”
$datesult | Add-Content C:watchdogresult.txt
Save as script.ps1
and create a task scheduler task with parameters:
Run whether user is logged on or not
Run with highest privileges
Repeat every 5 minutes for 1 day
Start program: powershell
Add argument: -ExecutionPolicy bypass -file “C:watchdognet.ps1”
Your txt file should have similar entries – UP if network is working and Repaired with date and time if team failed and was repaired by scrpt:
02/01/2015 12:25:49 Repaired
In Windows server 2012 R2 we can find a new DR functionality – extended hyper-v replica. In Windows server 2012 Hyper-V – replica was introduced – but some parameters were not as flexible as they are in a 2012 R2 preview. There was also possiblity to make hyper-v replica only to one location – so virtual machine that was running on one host was replicated only to an additional hyper-v host. Now you are able to do so called extended replication. It means that you are now able to replicate from location 1 to location 2 and from location 2 to location 3. So it is not possible to send replicas directly from a first hyper-v host to two others but from first to second and from second to third.
To see how it works you can check this video: http://screencast.com/t/8ZdQwdh3CM
While we are waiting for Windows server 2012 R2 I just played with a feature that will come very useful after upgrading Hyper-V hosts to new version (R2). Microsoft did a great job with the possibility to live migrate (by using shared nothing live migration feature that was already available in Windows server 2012) from old Windows server 2012 hosts with Hyper-V to new Windows server 2012 R2 hosts with Hyper-V. So you will have zero downtime while upgrading your virtualization platform. This goes one way only – so only from older (Windows server 2012) to Windows server 2012 R2 and NOT vice versa.
Here is a video how it works – at the end I also demonstrated that it does not work in opposite direction.
Update: It will fail if you have different name for virtual switch – If you have a new name for virtual switch on new server I suggest that you create an “fictive” private virtual switch with the same name on original (old hyper-v). You are not able to choose virtual switch on destination.