NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Horrible intermittent latency on the orbi router
Hello everyone,
I have had my orbi for a year now, and have loved every minute of it! Aside from a few intial issues with some of my smart home devices. I was able to iron those out, and I have recommended the Orbi to all of my coworkers, friends and family. However, since the latest firmware update, I have been having absolutely horrible intermittent latency.
Doing a trace route I was able to see the latency originating from my own network on any pc or device I tested from. I have seen this latency spike across the network when pinging one device from another, and have even see this with only the router connected, and the satttelite and modem disconnected.
I have attempted to factory default both devices many times. I have downgraded the firmware, and reupdated both my router and sattelite to no avail.
I have power cycled these devices, toggled them between AP And router mode. all to no avail. Again, this latency is experienced even across network from device to device using only the router, when disconnected from the carrier modem.
Any ideas?
24 Replies
Removed as a duplicate post.
Oh and to add a bit more detail. I am seeing this latency whether each device is connected via ethernet or wireless or a combination of both.
Keep in mind, this is easy to replicate even when only connected to the router via ethernet removing the Spectrum modem from the picture pinging across the router from one device to another.
Latency spikes are topping 3000+ milliseconds
Arkkady wrote:
Latency spikes are topping 3000+ milliseconds
Could you explain more about how latency is being measured? I have set up a Windows machine to generate 5,000 ICMP (ping) packets to a number of other devices connected via router, satellite, WiFi, ethernet, etc. When WiFi is involved in any part of the connection, there are occasional "spikes", but the largest I have recorded is 200ms. I do not doubt that something is going on, and would like to replicate your tests on my equipment.
What Firmware is currently loaded?
What is the Mfr and model# of the ISP modem the NG router is connected too?Is the Orbi system operating in Router or AP mode?
What is the size of your home? Sq Ft?
What is the distance between the router and satellite(s)? 30 feet is recommended in between them to begin with depending upon building materials when wirelessly connected.What channels are you using? Auto? Try setting manual channel 1, 6 or 11 on 2.4Ghz and any unused channel on 5Ghz.
Any Wifi Neighbors near by? If so, how many?Try enabling Beamforming and MIMO(MIMO may or maynot be needed) and WMM. Under Advanced Tab/Advanced Settings/Wireless Settings
Try disabling the following and see:
Daisy Chain, Fast Roaming, IPv6 and Set 20/40Mhz Coexistence to 40Mhz only. Save settings and reboot the router and satellite(s).
Arkkady wrote:
Hello everyone,
I have had my orbi for a year now, and have loved every minute of it! Aside from a few intial issues with some of my smart home devices. I was able to iron those out, and I have recommended the Orbi to all of my coworkers, friends and family. However, since the latest firmware update, I have been having absolutely horrible intermittent latency.
Doing a trace route I was able to see the latency originating from my own network on any pc or device I tested from. I have seen this latency spike across the network when pinging one device from another, and have even see this with only the router connected, and the satttelite and modem disconnected.
I have attempted to factory default both devices many times. I have downgraded the firmware, and reupdated both my router and sattelite to no avail.
I have power cycled these devices, toggled them between AP And router mode. all to no avail. Again, this latency is experienced even across network from device to device using only the router, when disconnected from the carrier modem.
Any ideas?
Thank you both for your replies!
The Firmware was recently automatically updated to the latest 2.3.1.60 firmware on both Satellite and router.
The router has been tested in both AP and Router Mode, with the same results. At this time however, the router and satellite have been placed back into router mode.
The ISP modem is not connected to this network at this time, as this is currently being used for point to point communication within the home. ( due to the issues I suddenly started having)
The home is approximately 2700 sqft. The Satellite had been located approximately 40 ft from the router with no issues for over 12 months.
However during troubleshooting this was tested at various lengths, measured out for consistency,
Tested using ethernet back haul directly next to each other, and at 10 ft, 20 ft 25 ft , 30 ft, and 40 ft with the same results.
Additionally tested with wireless back haul measure out in the same increments: 10 ft, 20 ft 25 ft , 30 ft, and 40 ft with the same results.
I was able to replicate the results using just the router between wired device and wireless devices.
Several Computers where used for testing.
2 windows based machines running windows 10 both with Wireless cards and wired gigabit ethernet ports.
2 Linux Based machines with wireless and gigabit ethernet cards.
This allowed me to confirm the same latency spikes occur regardless of a mix between wired wireless windows only, Linux only, and a windows / Liinux mix.
Each test was completed on the windows machines using the cmd prompt for simple ping testing, and win mtr for dynamic trace route results. All showing latency spikes above 3000 ms.
On the linux machine, I used the natively supported MTR and ping tools within the OS to the same result in all scenarios.
I have attempted to use all optional channels available for both 2.5 and 5ghz channels, and have toggled all available settings, such as: AP/Router mode, Beam Forming, MIMI, Daisy Chain, Fast Roaming, and unchecked the 20/40 MHZ Coexistence box.
I have rebooted the router and satellite several times, I have pin hole reset for 60/90/120 seconds the router and satellite to no avail.
I have also downgraded the firmware on both the router and satellite, which prompted the auto update, and I am still experiencing latency spikes across the LAN when isolated from the carrier Modem.
As far as the Neighboring Wifi, I have not seen any new wifi networks set up in the area, and since I am in an "established neighborhood" I am surrounded by retirees.
While the carrier modem is not in the topography at the moment, I have had the carrier Spectrum come out and replace the modem several times.
This was done in the beginning as I had initially suspected the WAN connection to be causing the latency spikes as they were originally observed while a member of the house hold was gaming.
The modems used, all of which produced the same results where: Hitron, Technicolor, Cisco, and Ubee.
I am currently using the UBEE, When initially troubleshooting the carrier modem I ensured that each device had, Wireless, firewall, and SIP ALG disabled.
All signal levels on all carrier channels in the modem are within spec:
Currently:
Downstream SNR: 39 dB
Upstream SNR: 38 dB
Downstream Power: 0.1 dBmv ( Dang near perfect)
Upstream Power: 41 dBmv ( again excellent)
No T3 or T 4 timeouts observed.
As you can see I have been working on this for a bit, and have tried to be as systematic in my testing as possible.
A shame that one of the best routers I have ever seen has suddenly stopped working after 12 months of flawless operation.
Any help would be greatly appreciated.
Arkkady wrote:Each test was completed on the windows machines using the cmd prompt for simple ping testing, and win mtr for dynamic trace route results. All showing latency spikes above 3000 ms.
Thanks for the details... a LOT of effort. Some questions:
- What is the average (typical) ping response time
- How often do these 3,000ms responses occur? (every 5 minutes? once an hour?)
- Is there a regular pattern to the long responses?
- How many long responses in a row? (only one, several?)
- Are the ping tests all run from the same machine, using other computers for response?
I configured some ICMP tests between Windows & Linux machines, also connected via ethernet & WiFi (all 5G now that I think of it). Two Windows 7 desktops, one Windows 10 laptop, Linux Mint on a laptop, and Raspberry pi3b+. My favorite tool is "hrping" for Windows because it reports in fractions of ms (not terribly significant when one is observing 3,000ms). I typically generate 5,000 ICMP requests which takes about 83 miinutes. When any WiFi exists in the connection (to the device or satellite backhaul), responses are on average longer and maximum response time is higher. Everything wired is always faster.
In all those tests, ping responses are nearly always between 2 and 7ms. The longest responses were from the Windows 10 laptop over 5G WiFi at 116 and 218 ms, but even then 95% of those responses were under 12ms. In all of my testing, I never had a ping response over 250ms. Not once.
I am absolutely convinced that what you observe is indeed happening. Since it appears not to happen to me, there must be something different between our installations.
I just did 1,500 pings to 8.8.8.8 and every value was less than 35ms (yeah Spectrum!) So, there has to be something different between my Orbi and your Orbi. It is tempting to think there is some regular WiFi interference (microwave, phone, ???), but the latency is reported to occur even when two computers are attached to the Orbi router with ethernet cable. i.e. no internet, no WiFi. And, it doesn't matter which two computers.
I have a pretty plain network, about 25 WiFi devices and 5 ethernet devices. Actually, very little constant traffic.
One thing that used to drive us crazy at work was "ARP storms". We would mess up and create a "loop" where there was more than one network path to a device. An ARP request would go "round and round", gathering packets until it brought the whole network down. I guess what I'm looking for is some network pattern that would generate so many packets all at once that the Orbi would be overwhelmed and literally "pause". With this phenomenon happening as often as every 5 or 10 seconds, it might be useful to do a Wireshark capture of the LAN side and "see what's there". One way is to use the debug page to set up a debug capture of LAN/WAN traffic. Capture one or two minutes, and then look at it.
I thought it could be a DOS attack, Cam Table overflow, or perhaps arp poisoning, which prompted me to isolate the internal network. Already tried a Wireshark session. Nothing found except occasional packet-loss and dropped packets. Very very strange.
I have changed my modem from a static to dynamic configuration to no avail either.
Its the craziest thing, has anyone been able to force the Orbi to stop auto-updating? I would really like to see if the firmware update had any impact on my orbi specifically.
I had looked into thte debug page previously to see if perhaps the memory was being overused and nothing.
Looks like I spoke too soon. When I changed the modem config to dynamic from static, the latency when from completely horrendous to simply terrible. max spike from 3k+ to 1k so, thats something.,
- This is the WORST WORST WIRST router EVER!!! I have to reboot it constantly. I have spent innumerable hours trying to solve problems and accusing my cable company of providing poor service. But after three new modems and being able to hardwire perfectly fine, I KNOW it’s the router. But I can’t find out what’s going on because my “tech support” expired after 30 days. My advice to anyone who has this piece of crap, and you are within the return window...GET RID OF IT.
