This topic has been marked solved and closed to new posts due to inactivity. We hope you'll join the conversation by posting to an open topic or starting a new one.
I have two (2) Proxmox hosts. Each host is connected to my network (172.20.73.0) via a wireless NIC (I tried wired and had the same problem) and I am using NAT and MASQUERADE. On each Proxmox host, I have a vSwitch; one host as 172.16.77.0 and the other host as 192.168.77.0. This has been working fine for months on my old Orbi Mesh system (RBR50).
Since upgrading to the RBRE960, I have noticed connections to containers and VMs fail. I can be in via SSH or RDP for hours while at other times, the connection freezes after seconds or minutes. Most times I can reconnect immediately, while other times, I must wait minutes or hours, or, reboot the Proxmox host. I am connecting from my laptop which is connected to WiFi.
The containers and virtual machines can communicate with one another, as well as, externally to my 172.20.73.0 network and the Internet.
I included a trace route from my laptop at the end of this message.
I am puzzled and looking for ideas and suggestions on how to address this. Your help and feedback is much appreciated. Thank you!!
Code:
PS C:\Users\admin> tracert 172.16.77.2
Tracing route to 172.16.77.2 over a maximum of 30 hops
1 3 ms 3 ms 4 ms router.lab.local [172.20.73.1]
2 9 ms 5 ms 7 ms proxmox1.lab.local [172.20.73.11]
3 248 ms 215 ms 432 ms 172.16.77.2
Trace complete.
PS C:\Users\admin> tracert 172.16.77.25
Tracing route to 172.16.77.25 over a maximum of 30 hops
1 5 ms 4 ms 3 ms router.lab.local [172.20.73.1]
2 46 ms 6 ms 23 ms proxmox1.lab.local [172.20.73.11]
3 63 ms 8 ms 6 ms 172.16.77.25
Trace complete.
PS C:\Users\admin> tracert 172.16.77.25
Tracing route to 172.16.77.25 over a maximum of 30 hops
1 5 ms 3 ms 3 ms router.lab.local [172.20.73.1]
2 218 ms 23 ms 8 ms proxmox1.lab.local [172.20.73.11]
3 459 ms 8 ms 67 ms 172.16.77.25
Trace complete.
PS C:\Users\admin> tracert 172.16.77.2
If I understood support correctly, they are having a known issue with the RDP protocol on the RBRE series. Engineering is working on a fix; however, the agent was not sure of the ETA. The best guess is Q1 of 2024.
I am going to return this system and stick with the ASUS ZenWiFi Pro XT12. It has worked flawlessly, is much more feature rich (much better software), and, about half the price of the Orbi system.
I will say I do appreciate Netgear products and cannot speak negative about the company. I had great success with my RBR50 system, as well as, many other Netgear products over the years. Support and engineering from Netgear on this issue was quite good. I didn't spend much time in lower levels nor did I have to repeat my story repeatedly. The team worked very collaboratively with me and were always extremely polite and calm.
I did change from a 192.168.1.0 network to 172.20.73.0 network to avoid conflicting with more common networks (for when I VPN). Does Orbi get confused with 172 networks?
192.168 shouldn't be conflicting with any VPN networks. VPN should auto avoid and configure networking that's not on same subnet as 192.168.1.x. You can also try using 192.168.0.1 as well. What I use.
I'd try the default subnet setting on the 960 to see if it works if you had been using that before on your prior Orbi system.
172 should work as well as it's a private subnet for home use.
@bryan3301 wrote: I did change from a 192.168.1.0 network to 172.20.73.0 network to avoid conflicting with more common networks (for when I VPN). Does Orbi get confused with 172 networks?
That is a perfectly acceptable solution and the RBR960 (like the RBR50 before it) should have no problem with it*. The trace routes show the laptop going through the router to the Proxmox host and then to the virtual server on the Proxmox.
Is it correct to assume that the router has a static route defining 172.20.73.11 as the gateway to 172.216.77.x?
* There was a post several months ago where the user wanted to VPN into two different Orbi systems and was frustrated that the both Orbi's defined their VPN subnets the same and could not be configured. As long as the target VPN networks are different from 172.xxxx and each other, this is a fine solution.
Intermittent issues are the very devil to diagnose. It is natural to assume that "What changed?" was the router transition. It would be an enormous chore to swap back to the RBR50 system to validate that the problem (a) never occurs with the RBR50, and (b) can be reproduced at will with the RBR960.
@CrimpOn Thank you for the continued feedback; this one is a head scratcher for me as I feel it should work. After all, it worked on my old RBR50. Routing and configuration were all set the same (with the exception of the IP address change, SSID/password change, and, the RBR960 has two extra WiFi networks, one for IoT and one for dedicated 6E).
Does the metric value on the route on the Orbi matter? I have tried with 2, 3, 10, 12, 15 and it does not seem to make a difference (not that I thought it would).
I have Tailscale configured on a couple hosts on this network as my VPN. I have no problems staying connecting to the servers when on VPN and they have no problems staying connected to the outside world. I only seem to have the issue when using my laptop at home on the WiFi network (I haven't tried physically connecting my laptop via Ethernet yet). This leads me to think it is bug with static routing on Orbi.
ORBI STATIC ROUTE CONFIGURATION
Route 1
Name: lab1.local
Private: Checked (I also tried /w unchecked) Active: Checked Destination IP Address: 172.16.77.0
Subnet Mask: 255.255.255.0
Gateway IP Address: 172.20.73.11
Metric:10
Route 2 Name: lab2.local Private: Checked (I also tried /w unchecked) Active: Checked Destination IP Address: 192.168.77.0 Subnet Mask: 255.255.255.0 Gateway IP Address: 172.20.73.12 Metric:10
172.20.73.11/24 (172.16.77.0/24) and 172.20.73.12/24 (192.168.77.0/24) are the Proxmox hosts. In parenthesis is the vSwitch network for each.
I only have the 960 devices now. The router upstairs and two satellites on the main floor at opposite ends of the house (right/front and left/rear). All devices are sitting exactly where the old ones did.
The only wireless networks I have are the ones built-in to the Orbi. I'll try turning off some of the other networks to see if that helps (i.e.- the IoT network, the guest network, and the 6E network).
I tried turning off all the other networks (6E, IoT, & Guest) and turning off AX features. The issue persists. It is seeming like the RDP connection to the VM drops after TCP Spurious Retransmission.
It appears the problem lies with the satellites and/or the backhaul WiFi network. This morning I shut down my two satellites and only connected to the router via WiFi. So far, no connection drops.
Updates to my testing... This morning, I shut both satellites down and connected just to the router. I did not have any connection drops to my VM. I then turned one of the satellites on, made sure I was connected it it, and again did not have any connection drops. I powered that satellite off and repeated the test with the other satellite; again, no connection drops. I then powered the second satellite back on, and so far, no connection drops.
I am wondering if turning the satellites on one at a time, and letting the satellite fully connect before turning on the second one, resolved the issue.
I have 1 router and 2 satellites in about 3,000 square feet. The router is upstairs in the middle/front of the house, one satellite is downstairs in the right/front of the house, and the other satellite is downstairs in the left/rear of the house. This is likely a bit overkill for inside; however, I do this to support signal strength outside to my blink cameras.
So far, the connection to my VM is still rock solid. Before this morning, it would only last less than a minute to maybe a couple minutes.
Thank you for the suggestion. If the problems arise again, I will either turn off the right/front satellite, or, reduce it's power. I have less of a need for WiFi in that area of the outside yard.
Thank you for the suggestion. If the problems arise again, I will either turn off the right/front satellite, or, reduce it's power. I have less of a need for WiFi in that area of the outside yard.
The Transmit Power setting affects the user facing radios on all units. There is no capability to set each unit to a different power level.
The distance between Orbi units is a balancing issue.
Performance of the router to satellite backhaul link is directly affected by distance.
The closer units are, the higher the connection rate will be.
The farther away the units are, the lower the connection rate will be.
In practical terms, a distance of approximately 30 ft. produces a connection that is "good enough" to support most user activity. This is affected by building materials. In my case, Orbi's claiming a backhaul speed of 1,733Mbps actually achieve about 760Mbps when placed about 25 ft. apart but going through several walls and a wooden floor. Since my internet speed is less than that, user performance is not affected.
WiFi coverage is a function of the distance as well. If router and satellite were two feet apart, the backhaul link would be the maximum theoretical rate which is the number quoted in the Product Data Sheet. But, the total amount of are covered would be almost identical to the router's coverage by itself. i.e., "nothing gained". Placing devices approximately 30 ft. apart produces the most additional coverage while supporting good user performance.
Placing units farther apart also reduces the problem of individual WiFi devices "flipping constantly." WiFi signal levels are not constant. As some devices compare available WiFi access points over time, some devices are more likely to get into a confused state where they keep changing from one access point to another, then back again, over and over again. Reducing Transmit Power has the effect of making units that are farther apart seem much less attractive.
So the goal is to have everything good enough to be useable. Backhaul link strong enough. Coverage Area adequate.Confusion limited.
I spoke too soon. This is so odd. With ONLY my router on, and no satellites, my connection is dropping again. Earlier this afternoon, things were rock solid for at least an hour.
Does the front LED turn on PINK when the happens on the RBR?
What is the Mfr and model# of the Internet Service Providers modem/ONT the NG router is connected too? Be sure your using a good quality LAN cable between the modem and router. CAT6A STP is recommended.
My issues are not with internet-facing connectivity. I do not even have issues when connected via Ethernet. These issues are only when I connect via WiFi and SSH or RDP to my lab machines.
I did perform a factory reset after I updated the firmware. I am not using Orbi for DHCP or DNS.
I have had a support case open with Netgear and am hopefully getting on with them this afternoon. Hopefully they can review the debug logs and we can figure out what is going on and how to address. If all else fails, I am thinking of giving ASUS ZenWiFi Pro a try.
