Introduction to vCloud Hybrid Services (vCHS) – Disaster Recovery (DR)
I just recently received my credentials to VMware’s vCloud Hybrid Services. One of the first things I was interested in testing out was the Disaster Recovery Service, which uses VMware’s vSphere Replication technology to protect VMs in your on-premise DC to vCHS. The following post provides the steps to configure the replication target as your vCHS VDC (virtual data center), and then configuring replication on a VM.
Part 1 – Select Target
The first thing to note is that the vSphere Replication appliances used for disaster recovery with vCHS is not the standard vSphere Replication appliance. There is a special build for this service. In this case, the build I used is 1667021. However this is subject to change, and you should verify which is the correct build number to use for your environment via supported documentation.
This new build provides an additional item in the vSphere Replication Management of Target Sites view. Now, as well as the DC target, there is a new cloud target.When you click this cloud target (second icon), you are provided with a number of connection settings including the provider address and the org name. Now this will vary depending on which vCHS instance you are using for the service. So where do you get this required information? Well, you will need to login into your vCHS account and retrieve the vCloud Director API URL, which appears under the Related Links on the right hand side of your dashboard, as shown below.Click on the URL expands it, and this is the information you need to capture.
This information can now be used to populate the Cloud provider address and organization name. Of course, you will also need to provide your vCHS credentials at this point.Now you need to select your Virtual Datacenter. In this case, I only have one VDC in my organization.When the setup completes, your vCHS org will appear as a valid target site. You will note that there is an issue reported with the networks. This needs to be configured next, as an appropriate network is required in vCHS for both test failovers and actual recovery.Clicking on the network link will launch the configure target network wizard. A number of networks are already created by vCHS for your convenience in the org (although you can also create your own). One of the networks is isolated, the other is routed. I am going to pick the isolated network for the purposes of this post. Once the networks are created, the status of the target site changes to connected. That completes the target site setup. The next step is to configure replication on a per VM basis.
Part 2 Configure Replication
vSphere replication is configured on a per VM basis, and replication a VM to the cloud provides almost the same experience when compared to configuring replication across DCs. Right click on the VM, navigate down to vSphere Replication and select Configure Replication as shown below:
Note that there are two options now displayed; replicate to another vCenter/DC or replicate to vCHS/cloud provider. Obviously in this case we are going to replicate to a cloud provider:
Since we have already setup our connection, we can choose that VDC. If we had not yet set it up, or we wished to connect to a different VDC, then we could select the New Provider VDC and go through the steps we just went through in part 1 above.The next step is to select an appropriate storage policy. This is provided by vCHS. In this example, there is only a single storage policy available for me to select called DR-Standard. Note that you also have the option of using replication seeds. If there is an existing copy of your VM on vCHS, you can use that to speed up replication time. This is useful in the event of a recovery, where the VM is moved form vCHS back to your DC, and you wish to re-establish protection. In this case, the copy of the VM which is still up on vCHS can be used for seeding. Since this is an initial configuration, and there is no VM available for seeding, we will omit that from the selection:
Next step is Guest OS quiescing. MS Shadow Copy Service (VSS) is the only one available for selection.Now it is time to select the Recovery Point Objective or RPO. The lowest RPO time is 15 minutes, the same as vSphere Replication across vCenters. This can be extended to an RPO time of 24 hours, if you so wish. By default it is 4 hours, and this is what we will leave it at.When the wizard completes, replication will commence (initial sync). The status may be observed in the Summary tab of the VM.
Part 3 Verify Replication via vCHS
When the synchronization is taking place, it may also be observed from the vCHS side. Here is a view of the replication still in its initial full sync state:
And that’s it. Eventually the VM will synchronize and kept in sync as per its RPO setting. As you can see, if you are already familiar with vSphere Replication then the process is pretty intuitive. I will admit that having to go to vCHS to get the URL and org are not that intuitive – perhaps something in the UI to inform you of this could be useful, or indeed, simply login to vCHS via the wizard and have it automatically retrieve this information for you would be ideal.
As I get to use this product more and more, I’ll share further posts.
Hey Cormac,
Do you have any idea when existing vCloud SP’s will be able to make use of this to provide their own end-point for DR-as-a-Service. (i.e. not using vCHS)
Right now Zerto own this market and it would be fantastic if VMware could offer some competition!
Cheers,
Gavin.
vCHS is not my area of expertise Gavin – I just wanted to capture the end-user experience in setting up DR using vCHS as an end-point. However, let me ask some people who might be able to offer a response to your query.
From the vCHS team, still under discussion, no ETA. HTH.
Cormac – thanks for the followup!
As long as ‘DRaaS’ to SPs has been identified as a need, then I’m happy. There was a ton of foot-stomping at last year’s VMworld about how partners are the lifeblood of VMware – it’d be great to see some proof of that!
Cheers,
Gavin.
Hello Cormac,
1. Will I still be able to verify the status of the replication from the esxi host using SSH when replicating to the cloud ?
vim-cmd hbrsvc/ commands ?
2. Also to check the connectivity for initial/ongoing replication, are the ports 31031 and 44046 still used ?
Thank you
Fazal
vim-cmd stuff is still there. Not sure about the ports, but that stuff is probably in the docs.
Hi Fazal,
Normal Replication will still continue to use the same ports, where as the Replication To Cloud will use ports 10000-10010
How do you actually invoke DR into the service?
Do you need to contact VMware support to start them up?
One assumes you have to purchase it. Don’t know the details tbh. there’s probably some info on the vcloud.vmware.com site.
No, you just start them. No need for tickets or phone calls. There’s even an API to do it…
Hi, is there a way to compress and optimize traffic between datcenter source and vCHS ?