From UntangleWiki

Untangle Server User's Guide

WAN Failover

WAN Failover Description Page WAN Failover video demo WAN Failover Forums WAN Failover FAQs

About WAN Failover

WAN Failover is a premium service that works in conjunction with multiple WAN interfaces to assure that you maintain a path to the internet if a loss of connectivity occurs on one of your WAN connections. It is intended to allow your Untangle to use connectivity provided by multiple ISPs. You must install and configure multiple WAN connections before you can take advantage of WAN Failover's capabilities. If you have not already done so, instructions for doing that are available here.

You may also wish to consider using WAN Balancer in your network as well. This would allow you to be able to distribute connectivity across multiple WAN connections on a regular basis, rather than simply maintaining a second connection should failover be required.

PLEASE NOTE that WAN Failover does not support PPPoE connections.

Installation

WAN Failover installs in the same manner as all other Untangle server applications and services. Under Apps (at the left side of the Untangle client screen), click on WAN Failover. You may purchase it or simply try it using the free trial. Once installed, it will appear in the Services section of your Untangle rack. The figure below is how WAN Failover appears in your rack.

Figure, WAN Failover faceplate

Because WAN Failover is a service, it affects all normal traffic in your network. You may use WAN Balancer in your network under normal operating conditions, but if failover is required, all traffic will be routed across the remaining functional connection(s).

Configuration

Once WAN Failover is installed in your rack, click Settings to access the WAN Failover configuration, then click Rules. You will see the Failure Detection Rules that are defined. If this is your first time on this screen, no rules have been defined. Click the Add button to define a failover rule. The following configuration screen will be displayed.

You may configure how WAN Failover works for each WAN interface. Begin by choosing a WAN interface. You may wish to provide a description of the interface so that you know which one is which. WAN Failover is based on test criteria that you provide here. If the testing that you specify fails, the interface will be considered to be off line and all traffic from this interface will now be routed to the other WAN connection(s).
These are the configuration options:

• Testing Interval determines how often (in seconds) your specified test will be executed.
• Timeout is the maximum amount of time that may pass without receiving a response to your test.

You should make sure that you allow for enough time to pass if you have a poor connection to the internet, or a connection that often has long latency (delays) associated with it. You should also make sure that the value you choose here is less than what you have specified for Testing Interval.

• Failure Threshold is how many failures are acceptable during the testing period.
• Test Type is the specific method you will use to determine whether failover will be initiated. In all cases, responses to the test method are expected to verify internet connectivity.
  The test types are as follows.:
  • Ping - A simple network ping test to the IP address that you specify.
  • Address Resolution Protocol (ARP) - No configuration is necessary for this. The WAN interface will automatically attempt to contact with the default gateway.
  • DNS - UDP packets will be transmitted to the previously-defined upstream DNS server.
  • HTTP - HTTP protocol will be used in transmission of TCP packets to the URL that you specify.

When specifying targets for ping or HTTP tests, you should choose a location external to your local network, and not too far upstream. Each upstream connection increases the possibility of latency and/or network problems creating false failures. Once you have entered these settings, click Save to lock in your settings.

Top

Related Topics

• Multi-WAN
• WAN Balancer

Top

WAN Failover FAQs

Is WAN Failover a free service?

No, it is a premium subscription. It is included in Professional Package, or you can purchase it separately if you do not have the full Professional Package.

I have installed and configured WAN Failover and nothing is happening. What should I do?

Check your Multi-WAN setup to make sure that you've set this part up properly. Also check your WAN Failoerv settings.

What tests should I use for Failover?

Untangle provides four test methods. In each case, your Untangle server sends out data packets and waits for an expected reply, up to the maximum amount of time that you have specified. You can use a ping test, which sends a specified number of bytes to the IP address that you specify via ICMP. You can use ARP, which Untangle will use to communicate with your default gateway. You can use DNS, where Untangle sends data via UDP to the upstream DNS server that you have previously configured. Finally, you can use HTTP requests via TCP to a URL (or IP address) of your choosing.

Is a Ping test better than the HTTP test?

Yes and no. Ping tests are simpler and more straight forward than the HTTP test, but many network operators block ping requests because they can be used for Denial of Service attacks. In both cases, you should select IP addresses that are external to your network but relatively close to you. As the number of network hops increases, the chances of encountering a bad or slow link increases. When that happens, Untangle may interpret it as a network problem and report one of your WAN connections as failing.

I only have one internet connection. Why would I want WAN Failover?

With a single WAN connection, its obvious that you have no alternative if your internet connection fails. You can monitor the uptime of your ISP with WAN Failover by defining a rule that will log service interruptions. If downtime is hurting you financially, WAN Failover can help you document that rather inexpensively.

Does WAN Failover support PPPoE?

No. This protocol is not supported in a failover environment.