In my latest posts, I have covered how to create the vault and configure the basics of the Azure Recovery Services – On Premises to Azure. Now we need the final part in order to make it work: a recovery plan. A recovery plan gathers virtual machines into groups and specifies the order in which the groups fail over.
I can see that my initial replication was completed (and it took 9 hours in my poor 2Mb upload plan):
You can see the events on the Hyper-V server that hosts the machine:
Cool. Now, let’s try the recovery plan ear, like I do. However, you should check the steps here: http://azure.microsoft.com/en-us/documentation/articles/hyper-v-recovery-manager-configure-vault/
Seems simple, right?
Should we just try to failover to azure? I will first try to start the VM, make some configurations and see what happens.
Let’s start it ‘On prem’:
It seems to work fine:
If my network mapping is correct, I should get an IP on the other side, in the mapped network. (See previous article).
Let’s test it!
Well, for a test it seems you will need an extra network. Let’s do it:
Let Azure take care of the DNS part in this case:
Cool, now we have a network. Let’s hit it!
It seems to be moving…
It seems the machine started. Let’s take a peek:
There you go.
But funny enough, I can’t connect to it.
No endpoints. Let’s try and create one:
Now the Connect option is enabled. Let’s try it…not that simple. If the original machine didn’t have the option for RDP enabled, it won’t magically work. Let’s take a look.
I will enable it and for the sake of simplicity, disable the firewall.
After doing that, I will end the test, wait for the replication to work and will try again.
Now, I have waited more than 30 minutes, to make sure that my 15 minutes cycles were covered. You can also check on Hyper-V manager:
Let’s test the failover again. So, it worked but I couldn’t log in. The account Administrator was the only account I had and not only that, but it also had a silly unsupported password.
I have created a new Local Azure Admin user and will try the failover test again once replication finishes. Let’s go!
Cool! First try! 😉 I’m in! Just got this screen, since the machine was being replicated and just booted up from the VHD:
And I’ve got the IP in the right network:
Now, the final FailOver test:
I’m going with planned.
Wow, it just stopped my VM on Hyper-V:
And started a “Planned Fail-over”
On the Azure side:
And after 7 minutes or so:
And there is our machine:
No endpoint though, which can be a good thing.
Let’s try internally:
Ok! Now, do you think we should fail it back?
I think we should!
But let’s try to trick it:
I’ve created a file on the desktop:
I will commit it:
Now I can failover back:
Will pick this one:
Now, on the VMM side:
This is taking a little longer than I expected, since there were no major changes:
And after 46 minutes…
My VM is back On Premises:
So, summary of Lessons learned:
– Configure another admin account, other than Administrator, to manage the machine
– Don’t use silly/unsupported passwords for those accounts
-Add an Endpoint after failing over, in order to access it (assuming you won’t be able to route through your VPN connection)
Hope this helps!