BLUF: When on my secondary WAN connection, GoodCloud (on host device and others on the network) connects and disconnects repeatedly.
Flint 2 (GL-MT6000) on 4.8.3
I have notifications set when any of my devices disconnect from GoodCloud, and that usually works just fine. Recently, however, my primary WAN went down, so I’m on my backup and I started getting a ton of connected/disconnected notifications from GoodCloud.
My first thought was there was some connectivity issue with my secondary WAN, but it was rock solid for Zoom, browsing, WiFi calling, etc. Then i wondered if there was an issue with that connection and GoodCloud, but I changed the order/priority of my WANs and it’s now been hours with no GoodCloud disconnects. So that means the issue has to do with the failover logic itself.
I plan to post to the appropriate group for that feature, too, but since no other services seem impacted, I figured I’d check to see if anyone else ran into this.
We tested this scenario locally using Flint 2 running firmware 4.8.3 and were unable to reproduce the issue.
When the connection failed over from Primary WAN to Secondary WAN, GoodCloud only reported one offline event (due to the Primary WAN going down) and one online event (after the Secondary WAN restored connectivity).
From your description, it seems the your Primary WAN was repeatedly fluctuating between available and unavailable states.
This caused the device to alternate its GoodCloud connection between the Primary and Secondary WANs, resulting in continuous online/offline notifications.
After adjusting the WAN priority so that the device uses the more stable Secondary WAN, the repeated notifications stopped.
One bit that I def should have mentioned, is that the cycling isn’t regular. It’d happen sometimes a couple times in a 15 min period, and other times may go an hour or so.
I don’t disagree that this sounds like the primary WAN being selected momentarily, but what’s curious is that nothing else is impacted during this time, even multiple Zoom and Meet calls (where I’d expect a brief period of inactivity to be easily noticed).
I’ll see if there is a daemon of some sort I could run to pull my IP every 30 sec for a while to observe for these switches - or see if the toggling of the WANs is logged somewhere on the router. Either way, however, if there is switching like that, what remains is that it seems there is less tolerance for instability on GoodCloud than on WiFi calling and Zoom…and that seems strange to me.
Is there some reason that GoodCloud would be so much more sensitive than even applications that have little room for any latency or instability?
