Diagnosing internet connection drops - help please

  • Release Candidate 6
    Guest:
    We are at a “proposed final” true release candidate with nothing known remaining to be changed or fixed. For the full story, please see this page in the "Pre-Release Announcements & Feedback" forum.
    /Steve.
  • Be sure to checkout “Tips & Tricks”
    Dear Guest Visitor → Once you register and log-in:

    This forum does not automatically send notices of new content. So if, for example, you would like to be notified by mail when Steve posts an update to his blog (or of any other specific activity anywhere else), you need to tell the system what to “Watch” for you. Please checkout the “Tips & Tricks” page for details about that... and other tips!

    /Steve.

squirrel

Active member
Oct 4, 2020
36
10
I’m a home user and don’t know anyone who would even understand my setup let alone be able to help diagnose the issue I’m having. I’m not in IT, mostly Steve has taught me what I know. A few years ago Steve said, “make up yourself a New Years resolution and get your devices isolated.” So I did. I bought a TP-Link TP1016DE 16 port managed switch, and two TP-Link access points (EAP 115). I cannot use a single wifi access point in my house as my house is long and made mostly of stone! The kitchen area is at one end and the siting room is at the other.

I’ve programmed the TP1016DE to have different VLANs for iots, guests and (devices, NAS and HP printer access). I‘ve set the DNS router to prevent all access to the internet from the printer. I figured that if the printer can’t see the internet, and the internet can’t see the printer, I’m pretty safe not bothering with any updates that may be available for it.

My setup is internet to the ISP provided NAT router to which my TV and security campers are connected; then on to a Synology NAT router which has packet inspection against Google’s and IBM‘s IP blacklists; then on to the managed switch; and then to the two EAP115 access points.

My problem is that I can’t think of a good way to identify the dropouts I get on wifi. I’ll be looking at YouTube or Netflix on my iPad, then after 10-30 mins, just sometimes, I get a freeze. Mostly I can fix the problem by going settings, toggle airplane mode, and after a few seconds normal service is restored. Today, it wasn’t restored straight away, which is unusual.

I don’t have a windows machine, but I have a Linux Mint machine and a Mac Book, and a variety of iOS devices.

I‘ve never had spooling problems using our TV when streaming, the TV is connected directly to the ISPs router. The drop outs last for 2-5 minutes. When the problem happened again today while watching YouTube on my iPad, I quickly went to settings and connected to the Synology router’s wifi and ran a download speed check, it was the max possible speed of 70Mbps. I swapped back to the EAP115 SSID and still no internet access, so I used the routers SSID for a while. 10 minutes later I tried the EAP115 SSID once again and it’s fine! It’s been fine all day! Sometimes this hang problem happens on a Firestick which uses an IoT SSID broadcast from the kitchen EAP115.

So the problem is not from the ISPs router out to the internet, that I’m sure of. It might be the Synology RT2600ac business router, may be it’s DNS is playing up and when I connect to its own SSID it allows access but not via it’s RJ45 input. May be it’s the TP1016DE which gets in a muddle, or the EAP115s. I haven’t yet moved from one EAP115 to the other when a hang is happening, I’ll try that next time it happen. If I reboot the switch internet access is restored but that may or may not be the source of the problem as toggling the airplane mode often solves the proble.

I tried connecting my Linux laptop to the secure VLAN by Ethernet for a week to see if the laptop‘s access to the internet dropped too, but it all ran properly for a week so I turned off the laptop and put it away In disgust.

I’m out of ideas to track down the issue by means that doesn’t disturb the connected devices which would then ask the router for a new DNS. By the way the Synology router provides all the IP addresses for the devices connected to the managed switch and I’ve allocated IP addresses to all my devices.

Thanks for any ideas you can offer.
 
identify the dropouts I get on wifi. I’ll be looking at YouTube or Netflix on my iPad, then after 10-30 mins, just sometimes, I get a freeze.
Well WiFi is not a wire, so it's shared. Is there something else in your area that is injecting noise into the wireless frequencies? A microwave would do it, as might any BlueTooth device (a wireless mouse or keyboard that is flakey, maybe.) Could just be bad firmware on the iPad... or maybe it's getting old?

One thing that may be happening is a fight between the two APs? The device gets confused by hearing both and can't decide who's in charge. To potentially help with this I would suggest you consider something mesh for your WiFi... There are a number of decent choices, but the Velop works well according to a friend... and can use a wired backhaul too.
 
Upvote 0
Well WiFi is not a wire, so it's shared. Is there something else in your area that is injecting noise into the wireless frequencies? A microwave would do it, as might any BlueTooth device (a wireless mouse or keyboard that is flakey, maybe.) Could just be bad firmware on the iPad... or maybe it's getting old?

One thing that may be happening is a fight between the two APs? The device gets confused by hearing both and can't decide who's in charge. To potentially help with this I would suggest you consider something mesh for your WiFi... There are a number of decent choices, but the Velop works well according to a friend... and can use a wired backhaul too.
Many thanks for your suggestions.
1. Interference: Microwave cookers. The last time I had the problem I was sitting in bed at 07:30. The nearest house is 40-50yds away, so I think that it’s not likely to be a cooker but something to put on the list while investigating other more likely causes.

2. Bluetooth: While sitting in bed upstair, it’s just my iPad 4 air and an iPhone 8 that’s on, and an iphone11 that’s asleep. No keyboard in use. I can’t see how it could be Bluetooth on this occasion.

3. AP fighting. Maybe. Next time it happens I’ll dash to the other AP and pull out its power lead as see what happen. Good one to try.

4. Mesh routing. From what I can figure out, when meshing I can’t have 3 separate VLANs, one for guests, one for IoTs, and one for my secure devices. And I don’t know whether they can isolate all devices from each other except my NAS and Printer. It’s really easy on a managed switch to configure which ports as assigned to which VLAN when connecting by Ethernet and which SSIDs of each AP gets assigned to each VLAN. I didn’t think that this level of allocation was possible using mesh routers.

5. I was wondering if the managed switch and/or APs get themselves in a muddle on occasion and a watchdog chip reboots them. I’ll dash to the switch an look at the lights and see if it’s blinking as usual or not.
 
Upvote 0
One quick thing you can do is allocate your different wireless access points to different channels.

2.4GHz: Use 1 and 11 (these channels are on opposite non-overlapping ends of the available spectrum)
5GHz: Choose 2 of these 3: 36-64, 96-132, 148-165

If your access points are fighting for airtime, this will at least reduce it from inside your network.
 
Upvote 0
I second the hard set channels for your two EAP wireless devices. You might hard set a "middle of the road" channel(s) for your other wireless device (Synology or ISP).

I'd check for firmware updates for your three TP-Link devices.

When the event happens again, I'd unplug the power cord from one of the EAP radio's/AP's and see if the issue resolves (could there be a conflict happening between your wireless device and the two AP's)?

When the event happens again, could you connect your device (that you're holding / that lost iNet connectivity) to one of your EAP based guest networks? Does the problem go away?

In cases such as this, I like to have each access point (EAP radio) set with a different SSID [Home-Sitting / Home-Kitchen] - that way, from your device, you can easily choose which access point to connect with. It would allow you to connect with the other EAP radio / AP, easily - and does the problem go away. (and, does the problem only happen when your device shows a connection to X SSID).

For now, aside from hard setting radio channels and checking for updates (firmware), much of my questions are pokes for additional information. Is it one of the EAP radios? Are the two EAP radio's fighting between each other?

When an outage happens, could you test a different iPad / wireless device - that is connected to the same SSID your problem device - does the other device have Internet access?

When an outage happens, could you test a different iPad / wireless device - that is connected to A DIFFERENT SSID - but one that is handled via your EAP AP's - does the other device have Internet access? Do all of the SSID's of the EAP AP's die or just one specific SSID?
 
Last edited:
  • Like
Reactions: squirrel
Upvote 0
So the problem is not from the ISPs router out to the internet, that I’m sure of. It might be the Synology RT2600ac business router, may be it’s DNS is playing up and when I connect to its own SSID it allows access but not via it’s RJ45 input. May be it’s the TP1016DE which gets in a muddle, or the EAP115s. I haven’t yet moved from one EAP115 to the other when a hang is happening, I’ll try that next time it happen. If I reboot the switch internet access is restored but that may or may not be the source of the problem as toggling the airplane mode often solves the proble.
When you lose connectivity, does it only happen with the device that you are using, or all devices connected via the EAP's/Synology? When was the Synology last rebooted? It sounds a bit like the DHCP table is getting confused. Are there any firmware updates for that?

You say that you have fixed addresses allocated for all your devices. Are they "reserved addresses" in the router, or static addresses set in the end device? Could there be any conflict?
 
Upvote 0
When you lose connectivity, does it only happen with the device that you are using, or all devices connected via the EAP's/Synology? When was the Synology last rebooted? It sounds a bit like the DHCP table is getting confused. Are there any firmware updates for that?

You say that you have fixed addresses allocated for all your devices. Are they "reserved addresses" in the router, or static addresses set in the end device? Could there be any conflict?
1. It appears that I can loose connection to the internet, but my wife’s device is fine, or the other way around.
A couple of weeks ago the kitchen amazon Firestick stopped accessing the internet and it was offline for at least 10 mins. We just switched over to terrestrial TV. Other devices still worked. I tried powering down the firestick and powering up again but that didn’t help.

2. The synology router is set to reboot itself everyday at 4am. Firmware updates go in automatically, I get notified so I know when an update has gone in. The last one was about 3 months ago. The local IP addresses used are reserved by a table in the router. I enter each device’s MAC address and set the IP I want it to use. No device demands a specific IP, no static addresses, they just get what is given to them. I don’t believe that there is any conflict. It not just one device that has a proble, it can be any using the wifi.
I had trouble getting my Mac on the wifi network a few weeks ago, I didn't have any time to investigate, so I used an Ethernet cable plugged into the managed switch allocated to the same secure VLAN that the wifi was trying to get on.
 
Upvote 0
Something more telling for the big issue - ping testing... Run this test when things are working to ensure what IP addresses to use and the expected results.

When you have a device that "looses Internet",
1) Does the device show an expected connection to a WiFi access point?
2) Does the device show an expected IP address?
3) Can the device ping a local IP address (access point, internal gateway)?
4) Can the device ping a nearby public address (the public IP for your network)?
5) Can the device ping a reliable Internet IP address (eg 1.1.1.1)?
Do the results show where a breakdown happens?


This comment is a very low burn thought:

Could you set a device or two - that you specifically use (for testing) - to not utilize a MAC address reserved IP address and let the device obtain an general IP via DHCP? Run that way for a week or two and see if other devices have outages and your test / DHCP devices not have issues.

Now, I don't see your MAC reserved IP addresses as a problem but the test could rule out your MAC reserved setup as the problem.
 
  • Like
Reactions: squirrel
Upvote 0
Something more telling for the big issue - ping testing... Run this test when things are working to ensure what IP addresses to use and the expected results.

When you have a device that "looses Internet",
1) Does the device show an expected connection to a WiFi access point?
2) Does the device show an expected IP address?
3) Can the device ping a local IP address (access point, internal gateway)?
4) Can the device ping a nearby public address (the public IP for your network)?
5) Can the device ping a reliable Internet IP address (eg 1.1.1.1)?
Do the results show where a breakdown happens?


This comment is a very low burn thought:

Could you set a device or two - that you specifically use (for testing) - to not utilize a MAC address reserved IP address and let the device obtain an general IP via DHCP? Run that way for a week or two and see if other devices have outages and your test / DHCP devices not have issues.

Now, I don't see your MAC reserved IP addresses as a problem but the test could rule out your MAC reserved setup as the problem.
Thank you for that list of suggested actions. I can answer some of them straight away but I will do as you suggest. I’ve also requested tech support from Synology to help who are usually very good indeed in resolving issues, I’ve been using their NAS‘s for years and they have been excellent.

1/2. Yes it does. Nice strong fan and the expected reserved IP address is shown in the info page on the device as expected.
3. I’ll try that.
4. No. Speedtest app starts off with a ping test and it can’t find anything.
5. No. But I’ll check again next time the dropout occurs.

I’ll try the unreserved device too.

Thank you.
 
  • Like
Reactions: JimWilliamson
Upvote 0
Not knowing the "guts" of Speedtest, I recommend a basic/clean/simple/reliable ping test. If your device does not have this option within the OS, find a suitable ping test specific app to tend the task. 0.02
 
  • Like
Reactions: squirrel
Upvote 0
I would suggest removing components to narrow down the problem. Or to go back to a previous configuration that was known to be working, and add one hardware component at a time to test when your connection problems start to occur.
 
  • Like
Reactions: squirrel
Upvote 0
Try one thing in case it is just this simple. Put something under the corners of the thing you think is failing. I have had 2 TP-Link devices in the past (I refuse to buy any more of their stuff), and I don't think they've hired anyone who knows anything about thermodynamics then or since. The products I had were too hot to pick up when they failed. But don't limit this to just TP-Link. Any device that just sits on furniture 24/7/365 may be overheating.
I have a wireless HDMI bridge (not from TPL) that would drop out on occasion. Two coasters were sufficient to fix it. A fan would be better.

Also, download a wifi scanner for your phone, check for the interference of neighboring wifi routers (just in case). On 2.4GHz, 10dB down may be marginally interfering, but 20dB down is not an issue at all. Probably need a bit more margin on 5GHz, and in the US it isn't channelized effectively (or at all).
Most scanners will do diagnostic pings to other devices on the LAN. SpeedTest is not a diagnostic tool. It has become an advertising tool for ISPs.

Apologies if this is a bit editorial.
 
Upvote 0
I think that I’ve found the source of the problem. The RT2600ac Synology router has a Threat Prevention feature. The feature checks every packet passing through it against a database maintained by Google and IBM. The database held on the router is updated daily and uses an SD card. When I checked if the Threat Prevention was still working it showed an error message saying it couldn’t build the database. I know that standard SD cards have a limited number of write cycles before they fail. I put in a new high endurance SD card and everything is now fine, stable and high speed functioning. I’ve had two standard SD cards fail this way I used for dashcams.

I reformatted the old card, and tried copying a 200MB file to it. It took just over 60 seconds. With the new card the file transfer was all over before I could start a timer.

Im not buying anything other that High Endurance SD cards from now on.
 
Upvote 0