|
Transparent failover to an alternate broadband network
Advanced console servers with mutiple broadband ports (such as dual Ethernet) can be configured to offer broadband failover. In the event of a disruption to the main network connection, access will switched transparently to the standby network connection. And when the main network connection is restored, access is switched back.
With this service enabled:
- the console server continually pings nominated probe addresses over the main network connection
- in the event of ping failure the console server switches to failover network connection
- the console server then is only accessible over this failover network connection. By default only SSH access is enabled on this failover connection. (However in firmware versions later than V3.0.1 HTTPS and SSH accees to the serial ports are also enabled)
- the administrator can then SSH (or HTTPS) connect to the console server and fix the problem
- in the event that the pings are successful, access switches back to the main network connection
You set up transparent failover when configuring the main Network 1 (eth0) connection in System: IP:
- Specify Network 2 (eth1) as the Failover Interface to be used when a fault has been detected with Network 1 (eth0).
- Then specify the Probe Addresses of two sites (the Primary and Secondary) that the console server is to ping over this network interface to determine if Network 1 (eth0) is operational
Next you configure Network 2 (eth1) as the Failover Interface so it is available as the transparent back-up port to Network 1 (eth0). Network 2 (eth1) will automatically and transparently take over the work of Network 1 (eth0) in the event Network 1 (eth0) becomes unavailable for any reason.
- On the Out-of-Band/Failover Interface tab uncheck Disable this network and configure Network 2 Settings (eth1) with the same IP setting that you used for Network 1 (eth0) to ensure transparent failover
By default, console servers with V3.1.0 firmware and later support automatic failure-recovery. So the console server continually pings probe addresses whilst in original and failover states. The original state will automatically be set as a priority and reestablished following three successful pings of the probe addresses (over the during failover. The failover state will be removed once the original state has been re-established.
Note: For firmware pre V3.1.0 the advanced console server does not support automatic failure-recovery back to the original state prior to the failover. So to restore networking to a recovered state the following command then needs to be run:
rm -f /var/run/*-failed-over && config -r ipconfig
If required, you can run a custom bash script when the device fails over. It is possible to use this script to implement automatic failure recovery, depending on your network setup. The script to create is:
/etc/config/scripts/interface-failover-alert
Failover and automatic failure recovery also can be configured to work with dial-out modems and ACM5000 3G connections. Refer faq 274
|
