There's nothing like that built in. What you probably want instead is a stretched/metro cluster where all your hosts in a cluster are in failure domains. This is a common deployment pattern with vSAN. Your main problem, if you decided to not do a stretched cluster and script it, will be this:
when the network of the SITE 2, for some reason, get down
How are you going to measure that? What determines when the network at site 2 is down? There's no easy trigger for that, so you'll be developing some automation and logic to determine how this gets measured.
What you are looking for is not a feature of vCenter itself. Instead, VMware offers Site Recovery Manager: https://www.vmware.com/products/site-recovery-manager.html
Detecting that a site went down etc. is almost impossible as there are a million+1 possible scenario's that can happen. Even products like SRM, need a human to press the "fail over" button. It's you, the admin, that must decide that a site is really dead and a fail-over is in order. There are so many things to consider. If a VM is restarted in the recovery-site, what about the networking aspects? How will clients reach the workloads that where running in the now defunct site? Designing a *working* fail-over construct can be quite a challenge and it's easy to overlook the little, but o so important things.
Also, a fail-over is a big deal, and so are fail-backs. Data needs to be re-synchronized between sites prior etc.
Thank you for answering.
What I mean when I say "get down" is: in case of power outage, a switch core malfunction or something like that.
srodenburg, In fact, I think this "Site Recovery Manager" is what I'm looking for. But, I don't know whether we have a licence or not about this product in the company I work for. I have to talk to my boss about it.
Thank you guys. I think my post ends here.
Here I am again,
I have another question about this subject: My boss said me something about "HA" and "ShadowSide"...is there something about these terms related to migration between wan sites in case of unavailability? As far as I know, HA has nothing to do with vms migrations between different sites, anyway...
Never heard the term "ShadowSide" before. HA only fails over VMs. If you have a stretched cluster, you can do things like failover to another site, but generally that's the work of a replication tool like SRM.
Thank you for answering. I googled something with your suggestions, and I ran into the following: vSAN Stretched Cluster, as shown in the links below:
Is that it what are you talking about? If so, can I configure it in my environment whitout SRM?
I'm sorry for asking so much questions...
No, it's not a replacement to SRM. Rather than guessing, why don't you explain what objective you're trying to reach here. What outcome are you expecting to have? And what is your current setup.
I'm sorry if I'm not being clear...let me try it again.
I have some important vms in my central office, in the Data Center (as showed in the picture I posted in my first post. In this case, the Site 2) that I need to keep them working properly in case of some kind of disaster, or unavailability, or something like that. I'd like these vms, in case of disaster, were migrated to another site (to Site 1, in the same picture). Something similar a Site Recovery, where the VMs keep working, no matter what happens.
Your need is very standard and an ideal use case for a replication tool like SRM or Veeam. What needs to change is you need each site managed by its own vCenter. The arrays I'm assuming are also separate, correct? The VMs from Site A would be replicated periodically to Site B. If an outage event happened at Site A, whatever replication solution you purchase will be able to, at your command, fail over the VMs and bring them up on Site B. It naturally isn't as simple as this what with networking considerations in mind, storage, distances, licensing, etc., but this is a very common request that is satisfied by such a solution. I'd recommend doing quite a bit of research on the Internet for architectures, whitepapers, blogs, documentation, etc to get a feel for the requirements and what to expect. If you or your team have limited knowledge and experience in this realm, it would be strongly recommended to find and work with a reputable VMware partner in your area who can guide you and implement the solution based on a proper design.
In short: In the current scenario, with only the VMware vCenter Enterprise Plus (this is the product we have in our environment), we cannot work with this type of solution, right? Can't use vSAN Stretched Cluster as a solution, right?
Correct, you need other solutions which require other licenses.