Testing Multi Wan

Started by sirio81, March 28, 2018, 11:01:22 AM

Previous topic - Next topic
Hi all, I successfully configured multi wan.
Now I'm playing with it to check if it works as I expected, and guess what...it doesn't  :D

My gateway group is named 'wangrp' and has 2 gateways
isp1 Tier 1 (this is te default gw)
isp2 Tier 2

On system / gateways / all I set 'Mark Gateway as Down' on the the gw isp1.
I check my ip by 'dig +short @resolver1.opendns.com myip.opendns.com' and also by sites like ping.eu and I still see the isp1 public ip.

I was expecting to see isp2 public ip.

I remove the flag 'Mark Gateway as Down' and aplly changes and check the publick ip.
Now both gateways are up so I expect to exit with isp1 but checking the publick ip, I see isp2 ip.
The exact opposite.
on Gateway / Status they are both online.

I repeated the procedure disabling the the isp1 gw instead of marking it down.
This time the connection is correctly router through isp2 gw.
I re-enable isp1 gw but I still get routed through isp2 gw (Tier 2) instead of isp gw (Tier 1).
Note: I disabled Sticky Connection' on firewall advanced settings.

Am I missing something?

At first you should test from internal hosts against speedtest providers if you have the correct IP.
Also it would be good to test failover in a real way like plug the cable or block a connection to the monitoring IP.

If this works we can work on dyndns ...

I checked by myip.com, ping.eu and other sites that shows my public ip.
I also checked form by linux pc querying opendns directly by command dig (no possible browser cache related problem there).

I also tested the real world scenario unplugging cables.

When I unplug isp1/Tier1 cable, both gateway go down, connections doesn't work off course.
When I unplug isp2/Tier2 cable, on the right gw goes down and connection works.

Note: "Trigger Level" is set to "packet loss or high latency".
I noticed it may behave differently if I set "Trigger Level" to "member down".
I'll report back after few more tests.

Sorry guys, but has anyone been testing multi wan on opnsense 18.1?
Just to understand if I'm the only one having troubles!

I made a factory restore and started form scratch.
My configuration, as describe before, is a gateway group with 2 gw:
  isp1 Tier 1 (also default gw)
  isp2 Tier 2.
  Trigger Level = Member Down

1) I mark down the gateway Tier 1:
    the isp1 gw status dosn't change (stay online).
   I get the same behaviour if I unplug the cable.

2) I mark down the gateway Tier 1:
    the isp2 gw status dosn't change (stay online).
   I get the same behaviour if I unplug the cable.

3) If I disable the isp1 gw, the connection gets corretly switched.

Please, may anyone test this "simple" scenario with 2 internet connection in fail over configuration.

Note: I also took care to set monitor ip different from the dns i set in system / settings / general.
Note 2: it behave the same also with "Trigger Level" Packet Loss.

I think the problem i repeatable.
I would like to open a bug report but it would be great if someone else can confirm it first.

You dont have to file a bug and I think I've tested multi wan more then any other ..

https://github.com/opnsense/core/issues/1821#issuecomment-370360822

Damn, I woke up this morning and I remembered i probably forgot to remove the flag "Disable Gateway Monitoring" after the re-configuration.
Today I'm not at work and I can't test it.
I'll do it the next week.
Thank you very much mimugmail for the support.

I'll soon jump in as well the multi-wan testing (on a regular basis).
I stupidly insist doing this with a second Miktrotik + a Huawei E3131 modem, for which I need an external antenna to stabilize the signal a little :)
Hopefully, I'll get one today.
OPNsense v18 | HW: Gigabyte Z370N-WIFI, i3-8100, 8GB RAM, 60GB SSD, | Controllers: 82575GB-quad, 82574, I221, I219-V | PPPoE: RDS Romania | Down: 980Mbit/s | Up: 500Mbit/s

Team Rebellion Member