Who Me Too'd this topic

t_k
Luminary
Luminary

Steps to reproduce the Orbi backhaul issues

My post about what you can expect with your Orbi system has received a good amount of activity. Mostly from users experiencing the same issues as I did and a few from users who felt compelled to post a vairant of "works for me!". 

 

Anyway, the Orbi backhaul issues are what make the system unusable and they are the most difficult to track down. Knowing this, Netgear is silent on the issue and just tries to turn up the volume on the parts of their product that do well (e.g. speed, design). Clearly, they are trying to drown out the noise from their user base that is wasting hundreds of hours trying make their wifi stable.

 

Below are steps to reproduce the backhaul issues at home. For folks who are working with Netgear support and getting stonewalled, you can use this to show the issue. I did and it worked to get a refund. For folks who aren't sure if you are also experiencing the issue, this will allow you the opportunity to conduct a specific test to see.  Since I don't believe the issue to be hardware, I'm confident everyone will see the same reults. Though, running the test is a PITA.

 


To be clear, this is not the only way to reproduce the backhaul issues. It's just the most reliable way I've come up with. The backhaul issues generally show themselves within a few minutes to a few hours - never more than 3 hours in my 2 satellite setup.

 

Prerequisits

  • An Orbi setup. Any number of satellites, 1 or greater
  • One system capable of running VMs on a bridged network per Orbi device (I used Docker, but any VM software will do)
  • One mobile device, capable of running ping (e.g. an smart phone or laptop)

 

Test Steps

  1. Connect each system capable of running VMs onto a corresponding Orbi, ideally via Ethernet so there is no chance of roaming. As a reminder, this is a WDS/backhaul issue, so the radios from the Orbi devices to various stations aren't involved in the issue.
  2. For each system capable of running a VM, spin up a few VM's with bridging configured on the interface connected to the Orbi (so fake MAC addresses generated and published through the backhaul). For reference, I generally created 3 on each.
  3. Install iPerf3 on a single VM on each system attached to the various Orbi's. Start an iPerf3 instance in server mode for each other Orbi device. So, if you have one router and two sattelites (three devices), you would start two iPerf3 server instances per system.
  4.  Run a continious iPerf3 test at some low bandwidth (say 20Mbps) from each system in step one to all the others.
  5. Now, periodically try to ping all the VMs while moving between Orbi's with the mobile device.

 

What you can expect

After some amount of time between just a minute and a few hours, you will no longer be able to ping some of your VMs. The behavior will change as you move between access points. Sometimes, the behavior will change just by waiting for 15-20 minutes.

 

This is the painfully difficult test I've ran tons of times with the same results.

 

While the test does sound complicated, it's actually using far less devices and activity than what you'll find in the real world. I suspect most Orbi Users have 15-30 real devices connected, and downloading can happen at an excess of 200Mbps - far more than this tests.

 

Last, I should state two things:

 

1.  I shared this test - though not written out exactly this way - with Netgear 6 months ago. I'm certain they never ran it. 

2. This stame test with a different wifi/WDS system, Ubiquity Unifi's, did not have any issues after 10+ hours.

Model: RBK53| Orbi Router + 2 Satellites Orbi WiFi System
Who Me Too'd this topic