How to Set Up Failover

A FlightVector Server can be configured to continue providing service if a single application server stops running. This is referred to as high availability or automatic failover. These instructions configure the FlightVector application server failover layer; SQL Server availability must be handled separately.

Instructions

1. Choose and implement a SQL Server high availability setup. This is not strictly required for one FlightVector application server to fail over to another, but SQL Server remains a single point of failure unless it also has a failover design. A common choice is a SQL Server Always On Availability Group. Configure the FlightVector ODBC connection to use the SQL listener, cluster name, or other supported SQL HA endpoint. Instructions are linked here as well.
    
2. Install the FlightVector Server on multiple machines, typically one primary application server and one failover application server.
   1. Make sure TCP traffic on port 26001 is allowed from client machines to each FlightVector Server. When automatic failover is enabled, port 26001 must also be reachable between the FlightVector Server machines.
   2. Make sure the same version of FlightVector Server is installed on each machine.
       
3. For each FlightVector Server install, clients and the other FlightVector Servers need a stable way to reach that server, such as a machine name, DNS name, or IP address.
   1. This is specified in the Settings.ini file in each server install folder, by default located at C:\Program Files (x86)\SofttechLLC\FlightVectorServer\Settings.ini.

      1. In each server's Settings.ini, add a FailoverHost value that identifies that server itself. Use a network name, DNS name, or IP address reachable by clients and by the other FlightVector Servers. 

         Example primary production server Settings.ini:

                     FailoverHost=PROD-FVSERVER

Example failover server Settings.ini:

                     FailoverHost=FAILOVER-FVSERVER

1. If copying Settings.ini from production to another server, do not copy FailoverIdentifier. Each server must have its own unique FailoverIdentifier. If it is missing, FVServer will generate it automatically when the service starts.
2. When each FVServer starts, it reads FailoverIdentifier and FailoverHost from Settings.ini and records or updates itself in the FailoverServer database table. The server returns the other active FailoverServer entries to clients as failover hosts during login.
3. Depending on your SQL Server configuration, make sure the ODBC connection string points to the SQL high availability group, SQL listener, cluster, or other SQL failover endpoint. 
4. Start the intended primary FVServer first if it is not already running. This is usually a Windows service called FVServer. 
5. In the FlightVector Settings menu, turn on "Enable Automatic Failover".
6. Start the failover FVServer service.

FailoverHost should not be the same value on every server. It should be the address of the server whose Settings.ini file you are editing.

If SQL Server high availability is not implemented, FlightVector application failover can still protect against an application server outage, but it will not protect against a SQL Server outage.