Intermittent network connectivity issues
Intermittent network connectivity issues
Hi all,
I recently bought an Omada setup but I am experiencing problems with my network connectivity. I found many forum posts with problems that look similar to mine. I have read them all and tried many suggested solutions, but for me it isn't solved yet.
I hope together we can pinpoint and solve the problems. Since I don't know exactly what causes it, I'll give you as much information as possible.
Network setup:
- ISP modem in bridge mode
- Asus RT-AC66U router with wifi disabled
- TP-link SG-108PEv2 switch
- OC200 cloud controller
- 2x EAP245v3 AP
- EAP225 outdoor AP
Furthermore:
- This system covers a household with one access point per floor. The amount of wireless devices is rather low (2 smartphones, 3 laptops, tv, chromecast, that's about it).
- All devices in the network are fully updated
- The wifi neighborhood is not very busy here. All my AP's typically show a utilization of less than 20%
- These problems never occured when I was using the router's build-in wifi, and also not over wired ethernet.
Problem description:
Regularly some devices in my network lose network connectivity, while claiming to still be connected to the wifi with full signal strength. This happens to multiple laptops and smartphones.
When this happens, the following is the case:
- The device can't ping anything but itself
- It still shows up in the clients list in the Omada controller
- Other devices which (according to the Omada controller) are connected to the same ssid, same frequency and same AP, keep on working without problems.
- Networking functionality is only restored after switching off and on the wifi, either from the device or from the AP
Turning all advanced features (fast roaming, airtime fairness, band steering...) off does improve the situation for some devices, but even then the problem still exists.
Who can help me to figure out the next step in solving this?
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
@Lucas_IT, check network routes next time the device loses contact.
Devices using policy routing or STP such as MacBooks often »lose« the default gateway after some time when connected to UBNT routers which resolve their hostname to 127.0.0.1. Maybe your Asus router does similar weird stuff.
- Copy Link
- Report Inappropriate Content
Hi @Lucas_IT,
Hmmm, that is strange indeed. How are your AP's connected to the router? Hardwired through the TP-Link Ethernet switch you have (perhaps I missed this in your post). How are they powered? Through the switch or via midspan injector?
I have a similar setup at home (except 2x EAP225) with about 38-45 wireless devices on any given day. My RF landscape is very crowded, but I don't generally get about ~15% utilization on my AP's radios unless I am active running a speed test or large down load on one of my devices simultaneously.
I'm running fixed channels, 3 & 9 for 2.4GHz (1,6,11 are very crowded), and 36 & 149 for 5.8GHz (20/40/80). There is a minor bug in the 2.7.0 firmware (for the 225's at least) where speeds will fall from ~400-500 Mbit/sec per 2x2 client to ~200 Mbit/sec for some unknown reason (TP-Link is looking into this) but not the loss of communications you have described. Is your 225-Outdoor also hardwired? Or are you trying to MESH it with the 245 Beta firmware?
How close in proximity are the two AP's. I know you said one per floor, but are the directly under/above one another? What is your transmit power. Definately try lowering it to Medium.
For now I would use separate SSID's for 2.4 and 5.8 so you can test your client connectivity one radio at a time. Does your ISP modem also have a router in it -- I assume, yes, since you say it is in bridging mode. You might try temporarily removing the Asus from the equation and re-enabling routing on your ISP modem and see if you still see the issue.
Be sure to use inSSIDER on one or more of your laptops to view the Wifi landscape and monitor signal levels.
It could be something as simple as a wonky Ethernet cable for example. If your devices are associated to one of the AP's and the AP's backhaul connection is flaky (poortly terminated cable, kink, etc.) the clients won't know that the AP and lost communication and will stay associated. You might have to try testing with one AP at a time to try and see if you can isolate the issue to a single AP/wiring connection
-Jonathan
- Copy Link
- Report Inappropriate Content
Hi @R1D2 and @JSchnee21,
Thanks for your quick and thorough replies.
After you mentioned it, I noticed for one of my Linux laptops that the default route in the routing table was set to 0.0.0.0 after the fault occured. Switching the wifi off and on restored it to it's proper setting. For my Windows laptop I also checked this after it lost connectivity, but there the routing table was stil sound.
Next I'm replacing the router with an old one that I still have lying around (loaded with an up-to-date FreshTomato firmware) to see whether it results in a more stable network.
If the problem persists I will indeed try to narrow down and pinpoint the problem like you suggested.
Answers to some of the questions:
- The AP's are all hardwired. The mesh function is disabled. One of the EAP245 is connected directly to the router, with the provided POE injector, while the other two AP's connect to the 802.11af switch.
- The AP's are technically only a few meters apart, but each floor is separated by a reinforced concrete floor. Transmit power is set to high, but that can probably be lowered.
- Currently the SSID's for 2.4 and 5 Ghz are combined, and the channels set to automatic. If needed, i'll try changing this.
Thanks for helping me out and I'll keep you updated!
- Copy Link
- Report Inappropriate Content
Hi @Lucas_IT,
Yes, I would manually set the AP's to different channels and lower the transmit powers to Medium. My AP's seem to want to automatically set temselves to the same channels for some strange reason.
I would also use separate SSID's for each radio as not all STA's work well with band steering. Though I've recent begun limited testing of band steering again in my house (using my iPhone XSM) and it seems to be working so far. I was hoping this would give me better range outside. But my coverage outside is still pretty poor and choppy. Decided to add one or two 225-Outdoors instead.
Aren't you're wireless clients usind DHCP? Wouldn't their gateway/router IP (e.g. 192.168.1.1 or whateve) be set as part of the DHCP lease?
-Jonathan
- Copy Link
- Report Inappropriate Content
Yes, the gateway is being sent with the rest of the DHCP information (that's also why everything works fine right after connecting), but it seems that in some cases, somehow this information is lost or overwritten in the clients. I need to do more research before I understand exactly what's going on.
- Copy Link
- Report Inappropriate Content
Lucas_IT wrote
After you mentioned it, I noticed for one of my Linux laptops that the default route in the routing table was set to 0.0.0.0 after the fault occured. Switching the wifi off and on restored it to it's proper setting. For my Windows laptop I also checked this after it lost connectivity, but there the routing table was stil sound.
If routes are lost out of the blue on some systems and other systems lose connectivity w/o losing routes, I think you should check the whole network setup.
- Copy Link
- Report Inappropriate Content
Hi all,
It's been a few days now since I have replaced my router with a new (old) one, and so far I have not experienced any of the problems. Both the Windows and the Linux machines I was talking about have a stable connection.
My best guess so far is that there's a bug in the Asus router which corrupts routes somehow, but which is only triggered in combination with the TP-Link gear, or some combination of different factors. When I have more time (unlikely to happen any time soon...) I will play with it some more to figure out the details. It might be an interesting and educative process.
For now I'm happy that everything works.
@R1D2 and @JSchnee21 , thanks for your help!
- Copy Link
- Report Inappropriate Content
- Copy Link
- Report Inappropriate Content
@Lucas_IT - I am having almost the same issues with my brand new TP-Link network. I have all new gear except my modem. I have a 600vpn router and a TP-Link easy smart switch. I am having issues with a couple of devices. One is connected to the network no problem but will not get the internet. the other is an iPhone 8 IOS device. Just drops the connection all together for no reason.
Like you, I never had these issues with either device when I was running on my junky modem/router/wifi device. It was a junk Netgear router I bought on sale at Walmart. I would have thought I would have a lot better network with all this TP-link hardware.
Any thoughts or help is appreciated.
Phil
- Copy Link
- Report Inappropriate Content
Hi Phil,
For the device(s) that keeps a connection but loses the internet, you can check the following:
- Can you still reach your LAN? And your gateway?
- Can you still ping the device from another place in your LAN?
- Is the routing table still correct?
- Check your DHCP server. Is there only one of those? Are the IP addresses of your devices as you expect them to be? Any conflicts with static adresses?
- Is the DNS set up correctly?
- Is the problem also present over wired ethernet?
Just some ideas. Maybe it helps.
For the device that drops the connection all together, that sounds more like a wifi problem. Please look at the suggestions that JSchnee21 made earlier in this topic.
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 4271
Replies: 11
Voters 0
No one has voted for it yet.