Mission statement
Rely on scaling a third data center automatically if and when data center one and two become unavailable.
Citrix GSLB
Load balancing citrix.company.com across three data centers hosting their own gateways and internal Storefront urls. Each Storefront internal url is of the set of servers physically located in that data center.
StoreFront servers set
A single group across the three data centers.
Storefront store
A core functionality of StoreFront is the ability to aggregate and de-duplicate common application and desktop resources from multiple Citrix Virtual Apps and Desktops (CVAD) Sites. This functionality is commonly referred to as multi-site aggregation.
Single store will provide all apps and desktops across the three data centers. The store will have two Citrix sites using site aggregation. Site one will have priority over site two.
Site: two
With this site having con's across all three data centers the site will be available at all times. Yet its applications and desktops will only be available, primarily, if site one becomes unavailable. Once the Storefront store checks against site one and two, if one is not found and all of two's apps and desktops will be available.
As users take up connections on vda's in site two other settings will providing for producing additional vda's to maintain a good user experience in real time. These features include but are not limited to Citrix: Autoscaling.
https://docs.citrix.com/en-us/citrix-daas/manage-deployment/autoscale.html
Continuous confirmation data center three is operating
By having con's in all three data centers we can have daily users working out of data center three vda's to understand and know with confidence data center three is fully functioning. This can be accomplished a number of ways. For example AD Group: DataCenter3Users. A small group of people can be added to this AD Group, which explicitly blocks access to App1 from site one. The result being they see App1 on site two.
Example one: business as usual
In this example all three data centers are healthy and available. Users sign in citrix.company.com and launches application: hyperspace. Results are application is launched from site one, through vda in data center one or two.
Business as usual
User is not presented with hyperspace from site two since site one is available. User has no possible way to reach site two's hyperspace application while site one is functioning.
Example two: data center one and two are unavailable
In this example data centers one and two have become unhealthy and unavailable. Users sign in citrix.company.com and launches application: hyperspace. Results are application is launched from site two, through vda in data center three.
Disaster recovery fail
User is not presented with hyperspace from site one since site one is unavailable. User has no possible way to reach site one's hyperspace application while site one is not functioning.
Through automatic scaling additional vda's will come online as needed to provide capacity for all users.
Final thoughts
As with anything there are more than one way to accomplish a mission. Understanding the what and why is going to be key. In this example we have accomplished providing a 3rd data center as a disaster recovery site, which we can confirm functions daily and have a high degree of confidence during a disaster we can recover.
