We just installed a Peplink Balance 20x and have Starlink as our main ISP, plus TMobile LTE SIM internally installed. Currently, there are some minor obstructions to our Starlink connection, so every few minutes the connection drops for a few seconds and then comes back online.
I have tried multiple configs to keep the failover time down to a minimum, i.e. roll over from WAN (Starlink) to TMobile, but there still seems to be a downtime between switching. Any recommendations on the best setting to enable so we have no downtime between LB failover?
I also tried “always on” for both, but the systems keeps defaulting to the LTE connection, and I want it to be on the WAN/Starlink one as main.
Speedfusion hotfailover ?
You want both networks as always on (priority 1).
Then you change the default “outbound policy” to Priority and put Starlink Above the LTE choice. (disable the HTTPS persistence rule)
For critical services you will want to use a SpeedFusion tunnel, either SF cloud or your own Fusionhub. You can search around for how to identify zoom, webex, gaming, VPN etc in the outbound policies, and route them via the tunnel.
3 Likes
Exactly what I was looking for. Worked like a charm. You are awesome!!!
This worked great for me as well. Thanks! The only issue that I have now is the WAN1 (starlink) failing the health test (DNS) every couple minutes. Using SpeedFusion I stay connected with WAN-EXT(t-mobile). My Starlink does go out periodically for obstructions but not as often as WAN1 fails the health check.
Hi. Are you not talking about a situation where health check fails because Starlink has? It would seem that the router is doing exactly what it is supposed to do. No?
Using SpeedFusion – good move!
No, the health check is failing more frequently than Starlink having obstructions.
OK, I’d note that there are other reasons why Starlink may present such failures, the chief of which is probably geometry – the constellation is not fully built-out yet. this is also why it works better at some latitudes than others. One thing you might do if you find the link generally reliable and it does not “chatter” up and down: You can reduce the “recovery retries” from 3 to 1.
We have peplink balance 30 running with Starlink old and new units (2nd generation) with the 2nd generation we had to overrule the WIFI wile using the ethernet adapter and from them it all works fine Healthcheck is on ping with dns 1.1.1.1 and 1.1.0.0
When you say overrule the wifi, do you mean that you selected bypass starlink router in the starlink app? I was wondering if I should have done that.
Yes bypass the wifi and continue with your own wifi trough the peplink.
1 Like
with us it works fine even on sea
1 Like
It looks like those of us with V2 dishes need to bypass the starlink wifi router or we get weird results like this. The V1 system allowed both the starlink and your own router to work concurrently.
starlink is not stable. it has lots of drops. the health check is likely operating as intended and you can tweak the health check settings. starlink is cool to have but far from reliable at this time.
It’s actually pretty stable here. The issue is with leaving starlink router enabled while using your ethernet port on another router. Needs to be in bypass mode. When the health checks are failing, other devices still connected to starlink router are working fine. But yes, it’s not the most stable solution so others may experience various other outages.
Hello-
How is your experience with the 2 Starlink WANs? Can you please provide some throughput/latency results? I am trying to do the same…whenever we initiate a SpeedFusion session our combined throughput speeds are significantly reduced from the SL connection (same with bonding 1 SL with 5G or bonding 1 SL with DSL, etc.). Any input appreciated!
Thanks
Following this - we are starting to play with two Starlinks and two+cellular. It all seems super dodgy so I wonder how people are setting up the balancing.
Even with the speedfusion cloud things seemed flaky, but I haven’t been able to really spend a lot of time testing (yet).set
Depends what you are trying to accomplish. If all wans are bad at the same time, you are out of luck. I needed three WANs to get a reliable Speedfusion Smoothing tunnel. Two WANs would be flakey for me, but adding a third flakey WAN, all three weren’t flakey at the same time.
1 Like
