There are several reasons that may cause this to occur:
- Using bad test points
- Testing is too sensitive
- Issues with the ISP’s WAN lines
How are bad test points fixed?
A test point is an IP address for a reliable host used by the Ecessa device to test for connectivity. By confirming connectivity to a remote network, the Ecessa can ensure that the WAN line is operational. It is important that the remote host must be able to reliably receive and respond to the test (either ICMP or TCP).
Commonly recommended testpoints include Google DNS servers (126.96.36.199 or 188.8.131.52) or OpenDNS IP addresses (184.108.40.206 or 220.127.116.11). However, any host that can be reached and responds reliably can be used as a testpoint.
The ping utility is a very useful tool to determine if a given test point is reliable. If connected to the Internet and unable to ping a test point, it is a bad test point (assuming the test point is configured for ICMP). If using TCP, connect to the test point on the configured port to confirm reliability. If the test fails, reconfigure the test point IP address to a different reliable host.
How is the sensitivity of the tests fixed?
Line Testing: This setting can be changed on the Configure WAN page under Advanced. Ecessa’s recommendation is to require one response out of three tests (1/3) per WAN line. If this value is set to fewer tests than three, it is considered sensitive. However, this setting can be changed to one response out of four tests (1/4) or one response out of 5 tests (1/5) if the WAN line is experiencing frequent test point failures.
Testing Interval (Recovery & Redundancy): This setting can be changed on the Configure Services page and determines the amount of time between tests. Increasing the value for this setting decreases the sensitivity of the WAN line. By default, the test interval is one second which is aggressive. The recommended test interval is three seconds. If the Recovery and Redundancy setting is disabled, all WAN lines will be reported as operational (UP status) by the Ecessa device. Disabling Recovery & Redundancy is not typically recommended as the Ecessa device is no longer testing the WAN lines so failures will not be reported.
Disable Recovery: This setting can be changed on the Configure WAN page under Advanced. This option disables line testing and recovery on the specified WAN. This setting will cause the Ecessa device to always report the line operational (UP status). This option is not typically recommended as the Ecessa device is no longer testing the WAN line so failures will not be reported.
Configure WAN > Advanced:
What if the ISP has a poorly performing line?
There are two options to compensate for a WAN line that performs poorly.
Time Out: This setting can be changed on the Configure WAN page. By default, the time out for test points is configured for one second. However, some lines may take over a second to receive replies from the test point. It is a rare occurrence but possible. It is recommended to increase the time outs to 1.5 seconds in this situation.
Configure WAN page:
Delayed Line Status Reporting: This setting can be changed on the Configure Services page. A WAN can experience the opposite problem when a test responds and falsely brings the WAN up and the intermittent link causes the Ecessa device to constantly change the line status. Enabling "Delayed Line Status Reporting" will delay line status changes to overcome this problem by waiting for the specified time configured under "Delay" seconds. If the link remains with the detected status after the specified time has passed, then the detected status will become the current status of the line. This setting only delays the line changing from DOWN to UP while the change from UP to DOWN is immediate.