WAN interfaces flapping with multiWAN

eri-- · Oct 21, 2008, 5:44 PM

Nothing has changed in that way apart the good thing that it is working now.
If you wish /usr/local/sbin/slbd.sh has the command to check the status. As far as i am conerned you may replace it with anything you please just return the status.

Maybe worth confirming that it is the syslog latency though 30sec are not that bad too :P.

cwadge · Oct 21, 2008, 11:20 PM

I'm discovering that on the latest nightly of 1.2.1rc1, the interface does get marked down after a few seconds. But unfortunately, the failover never actually takes places. If the interface goes down, it's simply down until you bring it back up. The downed network never falls back to the other gateway.

cwadge · Oct 22, 2008, 12:24 AM

@cwadge:

I'm discovering that on the latest nightly of 1.2.1rc1, the interface does get marked down after a few seconds. But unfortunately, the failover never actually takes places. If the interface goes down, it's simply down until you bring it back up. The downed network never falls back to the other gateway.

Ahhh, please disregard this. After some further troubleshooting, it turned out that the core router I was using to gauge uptime for that interface could be reached through either interface.

However, that does bring up two interesting points.

It took about 1 minute for the connection to completely fail over. The hardware is a dual 600MHz P3 with 384mb RAM, but it seems to be the loadbalancer that's taking so long to recognize that it's down, not the filter reload afterwords.
In this situation, I can only assume (until I dig a little deeper) that the PFSense box automatically started pinging the other gateway's core router through the gateway that was up. I'm not sure if this is considered a bug or just an undesirable feature… but ultimately, what's the difference?