Making Virtual IP Addresses Available to Applications in Citrix

Some applications, such as CRM and CTI, use an IP address for addressing, licensing, identification, or other purposes and thus require a unique IP address or a loopback address in sessions. Other applications may bind to a static port, which, because the port is already in use, causes the failure of multiple attempts to launch an application in a multiuser environment. For such applications to function correctly in a XenApp environment, a unique IP address is required for each device.

Use the virtual IP address feature to assign a static range of IP addresses to a server and have these addresses individually allocated to each session so that configured applications running within that session appear to have a unique address.

Processes require virtual IP if either:

• They use a hard-coded TCP port number, or
• They do both of the following:
  • Use Windows sockets, and
  • Require a unique IP address or require a specified TCP port number

Also, this feature lets you configure applications that depend on communication with localhost (127.0.0.1 by default) to use a unique virtual loopback address in the localhost range (127.*).
Processes require virtual loopback if either:

• They use the Windows socket loopback (localhost) address (127.0.0.1), or
• They use a hard-coded TCP port number

If the application requires an IP address for identification purposes only, configure your server to use the client IP address.

How Virtual IP Addressing Works

The virtual IP Address feature works as follows:

• During IMA startup, the virtual IP address assigner binds the assigned IP addresses to the NIC that matches the same subnet as the virtual addresses.
• When the virtual IP feature is enabled on a specific server, the virtual IP address allocator allocates all new sessions connecting to the server an address from   the  pool of available addresses that were assigned by the virtual IP address assigner.
• Each new session is allocated an address that is removed from the pool of available addresses. When the session logs off, the allocated address is returned to the available address pool.
• After an address is allocated to a session, it uses the allocated virtual address rather than the primary IP address for the system whenever the following calls are made:

  Bind¸closesocket¸connect, WSAConnect, WSAAccept, getpeername,
  getsockname, sendto, WSASendTo, WSASocketW, gethostbyname,
  gethostbyaddr, getnameinfo, getaddrinfo

Note: All processes that require this feature must be added to the Virtual IP Process list. Child processes do not inherit this functionality automatically. Processes can be configured with full paths or just the executable name. For security reasons, Citrix recommends that you use full paths.

 

To make virtual IP addresses available to aplications running in sessions

Use virtual IP addresses to provide published applications with unique IP addresses for use in sessions. This is especially important for Computer Telephony Integration (CTI) applications that are widely used in call centers.

Users of these applications can access them on a XenApp server in the same fashion that they access any other published application.

To assign virtual IP address ranges, you must have a reserved range of static IP addresses to assign to the server. Work with your network administrator to obtain a list of free addresses that are not part of your DHCP pool. Ensure that you do not include broadcast addresses.

Before assigning virtual IP address ranges, determine the maximum number of users you may have connecting concurrently to the server. Because every session connecting to the server is assigned an IP address (not just sessions launching the application that require virtual IP addresses), assign at least as many static IP addresses to the server as the maximum number of users who may be connecting concurrently to that server.

 

Note: In the event more sessions are launched on a server than IP addresses are available, the server displays the error message: “No virtual IP address is available for this session, please contact your administrator.” The inability of the server to assign a virtual IP address to a session does not prevent the user from launching an application that requires a virtual IP address within the session; however, the application may not function correctly.

• At the farm level, configure virtual IP address ranges and assign them to servers.
• Enable applications to use virtual IP addresses.

In addition to configuring virtual IP address ranges and enabling applications for use with virtual IP addresses, this feature can control and monitor virtual IP addresses available from each server.

 

To determine whether an application needs to use virtual IP addresses

Some applications cannot run in multiple sessions on XenApp. For example, if the application binds to a fixed TCP port on a specific IP address such as 0.0.0.0 or 127.0.0.1, this prevents multiple instances of the application from running in multiple sessions because the port is already in use. The virtual IP feature of XenApp can help solve this problem.

To determine whether or not the application needs to use virtual IP addresses:

1. Obtain the TCPView tool from Microsoft. This tool lists all applications that bind specific IP addresses and ports.
2. Disable the Resolve IP Addresses feature so that you see the addresses instead of host names.
3. Launch the application and, using TCPView, note which IP addresses and ports are opened by the application and which process names are opening these ports.

To use the virtual IP address feature, configure any processes that open the IP address of the server, 0.0.0.0, or 127.0.0.1. To ensure that an application does not open the same IP address on a different port, launch an additional instance of the application.