MWAN3 Advanced / Diagnostics buttons hanging

The MWAN3 LuCI page seems to have a minor bug: when I invoke buttons like Restart MWAN, Start MWAN (after a prior Stop), or Hotplug ifup, the page does not return normally. The “waiting for results” message continues for as long as I let it hang there. If I recycle the page, the page returns to “normal” - spinning dialogue message is gone, buttons enabled. I have not investigated - posting this in case others have seen similar behavior on that page. Would be useful to know if it’s not just my install.

1 Like

I ended up fixing this by writing a shell script to restart the wan interfaces. CL args allow me to select one or both interfaces; and the script runs a ping to verify ifup. Works, though I look forward to the GUI being fixed.

1 Like

Can I ask how you’re finding mwan3 on Turris at the moment? It hasn’t worked right for me for a month or more- both interfaces are connected and function correctly from the luci interface but my outgoing traffic is no longer multiplexed over both connections- and incoming only works on one WAN.

Hi Neil,

mwan3 has been reliable for me since February. The issues I continue to experience are caused by my unusual Internet connectivity (rural area with 4G LTE phones as our only access available - no wired anything except old fashioned telephone out here!!), not by mwan3 itself. From casual tracking of the traffic graphs, it appears to use both links; and falls back to the remaining link when one of the phones “leaves the premise”. Reconnect when a departed phone returns is problematic with our iPhone (an iOS / iPhone issue, again not mwan3), but is very quick with the Android.

The problem that you describe sounds like traffic policies out of sequence or disabled - or possibly rules out of whack. Here are my policies in the order that they appear on the LuCI page:

balanced
wan_m1_w3
wan2_m1_w2
unreachable (reject)

wan_wan2
wan_m1_w3
wan2_m2_w2
unreachable (reject)

wan2_wan
wan_m2_w3
wan2_m1_w2
unreachable (reject)

wan_only
wan_m1_w3
unreachable (reject)

wan2_only
wan2_m1_w2
unreachable (reject)

Here are the members, metrics and weights from my configuration. Note that wan1 is favored because there is off-of-network traffic on the device that serves wan2.

wan_m1_w3
wan
1
3

wan_m2_w3
wan
2
3

wan2_m1_w2
wan2
1
2

wan2_m2_w2
wan2
2
2

I had trouble briefly with a traffic rule that supported my access to an external vpn connection. I recall the issue blocked connection entirely, so different scenario. Worth checking rules, though, if you have added or changed any.

I will watch my outgoing traffic more closely to see if the distribution is actually what I thought I’d observed.

Robert

Any idea if there is any log showing how many times mwan3 moved to backup connection ?

I see only:

MWAN Interface Systemlog
No MWAN systemlog history found

Where for sure there was at least 1 switch.

Thank you.

To update you: MWAN3 simply started working for me after a Turris update. I therefore assume that whatever was breaking it was fixed.