I have a question, we have one main production site and two disaster recovery sites. helf load is in one Dr site while half is on other Dr site in case if main production site fails. Our management wants to use SRM. Is it possible in our case, i mean one main site and two recovery sites? SRM should recover half vms on one Dr site and half on other? mean one to many, one protection and two recovery sites?
thanks in advance.
Unless I'm missing something, the only design about multiple sites is the Shared Recovery Site (also called N-to-1), take a look: VMware KB: Setting up the shared recovery site feature in VMware Site Recovery Manager 4.x and 5.x
This is a good question and caused me to double check this as at first I thought that it surely must be supported but now I am not so sure.
Reading the operational limitations of SRM 5.8 there is no mention of a 1:N configuration where 1 is protected site and N is recovery site. So it would seem that it is not supported/possible.
I'll happily be told by basher or someone of equal prowess that I'm incorrect though
when it comes to SRM, Site pairing is always 1 to 1.
means even if you talk about N (Let's say multiple Remote/Branch Offices) - 1 (main office DC), I will have to install N number of instances at this main office DC to pair them with each individual remote/branch office. This is famous use-case we normally see is DRaaS.
now if we talk about 1 : N
where let's say I have an office, running with 100 VMs, I would like to protect 50 of those at Recovery Site A, and remaining 50 VMs at Recovery site B.
after sorting out replication of those VMs accordingly.
I would install two instances of SRM at main site, and have one instance of SRM to be installed at each recovery site.
Pair them, and carry on towards creating Protection Groups and Recovery Plans.
in this thread, one of the VMware employee reported that it's supported scenarios
First, I'd recommend reading this blog post by SRM PM BenMeadowcroft: VMware SRM Topologies | BenMeadowcroft.com
The 1:N topology is not explicitly mentioned, I guess it's covered under "Other".
With 1:N configuration you stand the risk of incidentally protecting one VM or Datastore using both SRM instances to both sites. Since every SRM is not aware of the others, you will get NO warning for this. If you hit "recover" on both SRMs you may end up with two copies of the same production VM, running in both recovery sites.