[SOLVED] Weird LAN/Wifi issue - must read

[Marcelmmx]

Hello,
In my house I am using a LAN as shown on this simplified drawing.
The primary router 192.168.0.1 is a Netgear D7000 VDSL router with the latest original firmware and the secondary 192.168.0.250 is a TP-Link TL-WR0146ND with DD-WRT v24SP2 (DHCP disabled) – connected via cable and used to improve WiFi coverage upstairs in my house. The switch is an unmanaged Netgear ProSafe 24 port gigabit switch.

When connecting using different wireless devices with both iOS and Android I have the following issue:
If the devices connect via WiFi A on my primary router, they can access the SAT receiver and the NAS without any issues.
When I connect them using WiFi B on my secondary router, both devices can access the SAT receiver and the NAS without any issues.

However, when I now switch back to WiFi A, neither device can access the SAT receiver and the NAS.
They don‘t see the devices on the network and cannot even ping them.
This situation lasts for some 8 minutes and 30 seconds after which the SAT and NAS become pingable and visible from my iOS and Android devices (this applies to all wired devices connected to my switch – all become visible after the said 8 minutes 30 seconds).

If I repeat this process – Wifi A – WiFi B – WiFi A the result is always the same and it takes 8.5 minutes till the SAT receiver and the NAS are visible on WiFi A again.

I have absolutely no idea what can be causing this. There are no other issues with the network, all wired and wireless devices are working properly, the speeds are good, the VDSL connection works fine. The issue now exists over 2 years and in that time the Netgear received several firmware updates, but nothing has changed.
Any advice appreciated.
Marcel

 

[Marcelmmx]

If you're using CAT 7 everywhere, from the wall plugs which computer devices plug into through the walls to the switch and from the switch to the router, that part is standardized.

BUT you are using different cable types from the computer devices to the wall plugs. I would replace those cables with the same CAT 7 scheme. Everything on the same cable type.

Be back later with info on the switch.

Ok, I will get cat7 patch cords and replace them all.

 

[Marcelmmx]

A little late here but I am curious about the IP addresses.

My understanding is:

- Router A (WIFI A) is the main router at 192.168.0.1 configured as the DHCP with the available IP DHCP IP address range being 192.168.0.2 to 192.168.0.199

- Router B (WIFI B) has been assigned a reserved static IP of 192.168.0.250 outside of Router's A available IP address range. DHCP disabled on Router B. ( Edit note: added underlined words to clarify.)

- SAT (being 192.168.0.200) and the NAS (being 192.168.0.201)

SAT and NAS using Static IP addresses and reserved by MAC on Router A or Router B? "Leftover"?

Where is the iPad's IP address of 192.168.0.205 coming from? Should be Router A (WIFI A) if I followed the preceding posts correctly. However, 192.168.0.205 is outside of the allowed DHCP IP address range.

Phone (Android) ?

May have missed something but I am still compelled to ask.

Wow, you sure read everything carefully
Sorry for that number, but that is a leftover from the time when I was experimenting with the setup and tested whether assigning a static IP to my iPad will change anything - it did not. Currently it uses a dynamic IP assigned by router A.
I also have several android phones and tablets connecting to my WiFi.

Marcel

 

[Marcelmmx]

Acknowledged. Researching.

Plug the NAS into the switch then unplug and replug in the switch and retest moving a wireless device from WIFI B to WIFI A twice to see if the problem remains afterwards while we research your switch specs and look for the problem.

I don't think plugging the NAS directly to the switch would help, because my SAT receiver is plugged directly into the switch and suffers the same problem - in fact the wireless device cannot see any wired devices until the switch resets (either by unplugging the power or waiting for 8 minutes).
In fact, I thought, that since the NAS does not go through the switch, rather directly to the router, it would be visible to my wifi devices at all times - but it isn't. I think this is one of the things that caused me to think the main router is the source of the problem and not the switch.

 

[LWFG001]

Netgear unmanaged Gigabyte JGS524 v2 switch data sheet: https://www.downloads.netgear.com/files/GDC/datasheet/en/JGS516v2-JGS524v2.pdf

Forwarding Rate for 1000 Mbps port: 1,488,000 frames/sec
Latency (using 64 byte packets) for 1000Mbps: 15μs (max)
RAM buffer: 256KB
MAC address database: 8K

Bottom line: Either this switch's RAM buffer is defective or it's something else because it should take this switch a split second not 8.5 minutes to discover those wireless devices when they pass from WiFi B to WiFi A. Let me explain.

Unmanaged switches are layer 2 devices which have no ip address or Ethernet MAC address. They do not store ip addresses; however, they do store MAC addresses and port information in RAM buffer volatile memory.

Note: Volatile memory requires continuous power to maintain what is stored there. This is why powering off and powering on the unmanaged switch resets the memory.

How they work is by examining the source and destination MAC addresses of an incoming frame. The source is saved in the MAC address table along with the receiving port in the unmanaged switch's RAM buffer memory, while the destination MAC address is looked up in the MAC address table to see if there is an associated port. If there is then the frame is forwarded out that port only, otherwise it gets broadcast out all ports except the source port (split horizon rule). During the whole process the frame remains unchanged.

An unmanaged switch just does this data exchange. It has firmware but it's not altered so not updated. There's no need for a web interface. There's no set-up involved beyond plugging it in, turning it on, and plugging cables into ports.

And because there's no way to access an unmanaged switch remotely, like you do with a managed switch, there's no way to see the mac address table and observe what's happening with the table unfortunately. If we could see it, we could track the wireless device as it moves between the routers on the switch and see if it's dropping.

So if the switch is the problem, the mac addresses of the wireless devices moving from WiFi B to WiFi A are dropping from the RAM buffer volatile memory in the switch because that volatile memory has become defective (it's a memory chip in the switch) forcing their rediscovery which is then also taking far too long because of the defective volatile memory.

This is a possibility but there's no indicator light for the RAM buffer to tell us if it's gone bad. There's just the power LED which should be on and the port light for each of the two ports which the two routers are plugged into which should be on and also blinking when activity is present.

The way to tell for sure if the memory chip in the switch hosting the RAM buffer with the MAC address table has gone bad and is the problem is to swap out your unmanaged switch with any other working unmanaged switch and see if the problem persists.

If the replacement unmanaged switch solves the problem, then that's what it was. However, on the other hand, if the replacement switch exhibits the same behavior taking 8.5 minutes when a wireless device changes from WiFi B to WiFi A then the Netgear unmanaged switch is fine and IT IS SOMETHING ELSE!

I'm not yelling. I'm just happy that we've finally isolated this switch issue to the point where we can know for sure. Swap the switch out and let us know if the problem persists. If so, it's not the switch and we'll keep looking.

In fact, while you're doing that, I'm going to take a look at your LAN ip configuration closely and see if that's the problem. I would have done that first but with all the carrying on in the thread about the switch I ended up side tracked. Well, no matter. Peace .

 

[Marcelmmx]

Netgear unmanaged Gigabyte JGS524 v2 switch data sheet: https://www.downloads.netgear.com/files/GDC/datasheet/en/JGS516v2-JGS524v2.pdf

Forwarding Rate for 1000 Mbps port: 1,488,000 frames/sec
Latency (using 64 byte packets) for 1000Mbps: 15μs (max)
RAM buffer: 256KB
MAC address database: 8K

Bottom line: Either this switch's RAM buffer is defective or it's something else because it should take this switch a split second not 8.5 minutes to discover those wireless devices when they pass from WiFi B to WiFi A. Let me explain.

Unmanaged switches are layer 2 devices which have no ip address or Ethernet MAC address. They do not store ip addresses; however, they do store MAC addresses and port information in RAM buffer volatile memory.

Note: Volatile memory requires continuous power to maintain what is stored there. This is why powering off and powering on the unmanaged switch resets the memory.

How they work is by examining the source and destination MAC addresses of an incoming frame. The source is saved in the MAC address table along with the receiving port in the unmanaged switch's RAM buffer memory, while the destination MAC address is looked up in the MAC address table to see if there is an associated port. If there is then the frame is forwarded out that port only, otherwise it gets broadcast out all ports except the source port (split horizon rule). During the whole process the frame remains unchanged.

An unmanaged switch just does this data exchange. It has firmware but it's not altered so not updated. There's no need for a web interface. There's no set-up involved beyond plugging it in, turning it on, and plugging cables into ports.

And because there's no way to access an unmanaged switch remotely, like you do with a managed switch, there's no way to see the mac address table and observe what's happening with the table unfortunately. If we could see it, we could track the wireless device as it moves between the routers on the switch and see if it's dropping.

So if the switch is the problem, the mac addresses of the wireless devices moving from WiFi B to WiFi A are dropping from the RAM buffer volatile memory in the switch because that volatile memory has become defective (it's a memory chip in the switch) forcing their rediscovery which is then also taking far too long because of the defective volatile memory.

This is a possibility but there's no indicator light for the RAM buffer to tell us if it's gone bad. There's just the power LED which should be on and the port light for each of the two ports which the two routers are plugged into which should be on and also blinking when activity is present.

The way to tell for sure if the memory chip in the switch hosting the RAM buffer with the MAC address table has gone bad and is the problem is to swap out your unmanaged switch with any other working unmanaged switch and see if the problem persists.

If the replacement unmanaged switch solves the problem, then that's what it was. However, on the other hand, if the replacement switch exhibits the same behavior taking 8.5 minutes when a wireless device changes from WiFi B to WiFi A then the Netgear unmanaged switch is fine and IT IS SOMETHING ELSE!

I'm not yelling. I'm just happy that we've finally isolated this switch issue to the point where we can know for sure. Swap the switch out and let us know if the problem persists. If so, it's not the switch and we'll keep looking.

In fact, while you're doing that, I'm going to take a look at your LAN ip configuration closely and see if that's the problem. I would have done that first but with all the carrying on in the thread about the switch I ended up side tracked. Well, no matter. Peace .

Wow, thank you so much for all the effort you put into this! I wish the Netgear support were at least half as useful as you

I will try to get another switch to test, and in the meantime, I will get the CAT7 patch cords.
As I was checking how many I need I have discovered that there is indeed one part of the "core" network that does not use CAT7 - Router B is connected using a patch cord that is CAT5e.
I should have the CAT7 cords tomorrow, so I will update then.
As for the LEDs on my switch - the power LED is on and all other LEDs for connected ports are blinking and correctly identifying devices which only use 100BASE devices. I specifically checked those for routers 1 and 2.
I also tried plugging routers 1 and 2 to different switch ports, but the issue remains unchanged.
Should you need some information regarding the IP config, let me know. (The iPad IP address on the drawing is not applicable anymore - it now gets a DHCP assigned one).

EDIT - found out the cabling is cat 6a, so ordering 6a patch cords. It's been almost 3 years, so I forgot.

Thank you once again.
Marcel

 

[Ralston18]

Wired connection question:

Router A,DHCP (Wifi A) [LAN] ---Ethernet ---> Switch ---Ethernet cable --->[LAN ]Router B (Wifi B)

Connection Switch to Router B is indeed to [LAN] and not to [WAN] - correct?

With regards to the 8.5 minutes:

If you reduce the allowed DHCP IP address range on Router A to just reflect the number of devices on your network with maybe 4 or 5 "extra" DHCP IP addresses, does the 8.5 minutes change?

I.e., DHCP range is currently 192.168.0.2 to 192.168.0.199.

What happens if you try 192.168.0.2 to 192.168.0.16 ("16" being arbitrarily selected on my part.)

And, after the range change, allow some time for all devices to get their tables updated.

 

[Marcelmmx]

Wired connection question:

Router A,DHCP (Wifi A) [LAN] ---Ethernet ---> Switch ---Ethernet cable --->[LAN ]Router B (Wifi B)

Connection Switch to Router B is indeed to [LAN] and not to [WAN] - correct?

With regards to the 8.5 minutes:

If you reduce the allowed DHCP IP address range on Router A to just reflect the number of devices on your network with maybe 4 or 5 "extra" DHCP IP addresses, does the 8.5 minutes change?

I.e., DHCP range is currently 192.168.0.2 to 192.168.0.199.

What happens if you try 192.168.0.2 to 192.168.0.16 ("16" being arbitrarily selected on my part.)

And, after the range change, allow some time for all devices to get their tables updated.

Hi,
yes, router B is connected to the switch via LAN - to the router's LAN (not WAN) port.
As for the address range, I changed it to .0.30 and nothing has changed timewise.
Marcel

 

[LWFG001]

Wow, thank you so much for all the effort you put into this! I wish the Netgear support were at least half as useful as you

I will try to get another switch to test, and in the meantime, I will get the CAT7 patch cords.
As I was checking how many I need I have discovered that there is indeed one part of the "core" network that does not use CAT7 - Router B is connected using a patch cord that is CAT5e.
I should have the CAT7 cords tomorrow, so I will update then.
As for the LEDs on my switch - the power LED is on and all other LEDs for connected ports are blinking and correctly identifying devices which only use 100BASE devices. I specifically checked those for routers 1 and 2.
I also tried plugging routers 1 and 2 to different switch ports, but the issue remains unchanged.
Should you need some information regarding the IP config, let me know. (The iPad IP address on the drawing is not applicable anymore - it now gets a DHCP assigned one).

EDIT - found out the cabling is cat 6a, so ordering 6a patch cords. It's been almost 3 years, so I forgot.

Thank you once again.
Marcel

Good job! Let me know if the problem persists after swapping the switch out. If it does persist with the new switch, it is NOT the switch. If the problem is fixed, it was the switch memory chip going bad.

And unmanaged switches don't work with nor store routing tables from route protocols like ARP so the person who posted that was wrong about it. Unmanaged switches only work with MAC addresses and port information. That's it.

When I get back from work, I'll start looking at your network IP configuration in case the new switch has the same problem. And, of course, if the new switch has the same problem then you can return it and get your money back because the old switch will have been shown to be good.

Also, make sure all the CAT cables on your network are the same type and your routers will love you. If you're on CAT 6a everywhere, that as a good thing. Look at each cable and visually inspect the markings say Cat 6A. Don't leave a straggler.

I did an earn an A+, CNE, and MSCE back in the day so am qualified to speak on networking.

 

[Ralston18]

Thanks.

8.5 minutes (510 seconds, 0.142 hours) must have some meaning..

Does the following link present the User Manual for Router A (Model D7000)?

https://www.downloads.netgear.com/files/GDC/D7000/D7000_UM_01July2015.pdf

My thought is that some clues may be found via the Router's logs - starting about Page 252.

Likely dependent on what activities are selected for monitoring.

 

[LWFG001]

Ralston18, pardon my gobbledygook but in this case I've started at the base of the ISO's OSI model and am moving up. We're making sure layer 1 (physical layer which is the cabling in this case) and layer 2 (data link layer which is the unmanaged switch in this case) are 100 percent before moving up to layers 3 and 4 where we look at IP and routing. By doing it this way, we're fixing problems across relevant layers to help make this man's home local area network 💯!

 

[Ralston18]

Nice way to work through it all.

I have been thinking about what might be discovered when all is working well vs what is happening during the 8.5 minutes.

Both ways: Wifi A to Wifi B, and then Wifi B to Wifi A.

Latency somewhere.

When you get to Layers 3 & 4 consider some way to observe your network and what is going on: netstat perhaps?

Wireshark?

In the meantime:

Router B is a TP-Link TL-0146TN - correct? Not finding any documentation.

 

[Marcelmmx]

Imp

Nice way to work through it all.

I have been thinking about what might be discovered when all is working well vs what is happening during the 8.5 minutes.

Both ways: Wifi A to Wifi B, and then Wifi B to Wifi A.

Latency somewhere.

When you get to Layers 3 & 4 consider some way to observe your network and what is going on: netstat perhaps?

Wireshark?

In the meantime:

Router B is a TP-Link TL-0146TN - correct? Not finding any documentation.

This is router B

 

[Ralston18]

Thank you.

I am going to take a look at the available configuration settings.

Noted that the Router B's status page is reporting 42 active connections - is that correct, expected, viable....?

 

[Marcelmmx]

Thank you.

I am going to take a look at the available configuration settings.

Noted that the Router B's status page is reporting 42 active connections - is that correct, expected, viable....?

I don't know what that number means, but after watching it for a while I saw it climb all the way to 125, then go back to 86 and then it just moves around 100.
On another note: here are some screenshots I took a year ago. I was pinging my sat receiver from my iPad. Look at the times. The ping was timing out till about 8 minutes from the WiFi switch, then some very long times began to appear, got gradually shorter till they finally normalised. Could this be indicative of some specific issue?
Screenshot 1
Screenshot 2
I did another time measurement today. It took the iPad over 9 minutes to connect to the sat receiver following the WiFi switch, so the 8:30 time is not carved in stone.
Marcel

 

[Ralston18]

Based on the label I would expect that the Active connections would be the number of devices connected to the network and I would not expect that number to be bouncing all around.

And I do not know what that symptom may indicate....

Two things come to mind: Something broadcasting or some network loop.

https://kb.netgear.com/000060475/What-is-a-network-loop

Going out of my comfort zone so hopefully someone may spot where I may be going astray.

192.168.0.200 - that is still the SAT receiver - correct?

Any more information on that SAT receiver: make, model, etc.?

Is there a User Manual/Guide with installation and configuration settings?

Do not want to start making changes to your plans per se; however, consider some testing without the SAT on the network.

 

[LWFG001]

While I'm waiting for result of swapping the switch out with another unmanaged switch to see if the problem persists, I want to pipe in here.

Ralston18 is correct that if you have a computer attached to the network on a CAT cable but also have wireless turned on and connected to the same network on the same computer: this results in a broadcast storm.

When you connect a computer or device to a network, connect it using either a CAT cable OR the wireless.

For a computer attached via a CAT cable, it's as easy as making sure that if it has wireless networking it's not enabled. Or, alternatively, if the computer is using wireless to connect to the network, then make sure the computer is not also on a CAT cable too.

 

[Marcelmmx]

Based on the label I would expect that the Active connections would be the number of devices connected to the network and I would not expect that number to be bouncing all around.

And I do not know what that symptom may indicate....

Two things come to mind: Something broadcasting or some network loop.

https://kb.netgear.com/000060475/What-is-a-network-loop

Going out of my comfort zone so hopefully someone may spot where I may be going astray.

192.168.0.200 - that is still the SAT receiver - correct?

Any more information on that SAT receiver: make, model, etc.?

Is there a User Manual/Guide with installation and configuration settings?

Do not want to start making changes to your plans per se; however, consider some testing without the SAT on the network.

As for the loop, I think the only possible way to create one in my case is the simultaneous use of WiFi and Ethernet - laptops, TV sets, Android box - all have both types of connection. Such device should be able to only choose one type of connection to prevent this kind of loop, shouldn't it? And it would show in the list of connected devices on the router twice, wouldn't it? Or should I disable all WiFi connections on wired devices?

Yes, the 200 IP is the sat receiver. VU+ Duo.
EDIT: I used a different tool today to ping my devices from my iPad and it took exactly 510 seconds to start receiving replies from my NAS and 390 seconds from my SAT receiver. There was no gradual "shortening of times" like on the screenshots - one ping timed out the next one was normal.

Marcel

 

[LWFG001]

As for the loop, I think the only possible way to create one in my case is the simultaneous use of WiFi and Ethernet - laptops, TV sets, Android box - all have both types of connection. Such device should be able to only choose one type of connection to prevent this kind of loop, shouldn't it?
Marcel

Yes. Operating systems have default priority levels set to decide which way to access a network in case multiple methods such as wireless and ethernet both exist on a computer/device, and/or multiple network interface cards are present on a computer/device which the operating system can choose from to connect to the network with.

Ethernet, by default, is set on operating systems to have priority over WiFi.

So if you have a notebook computer with wireless enabled and the notebook is connected to the network over WiFi, but then you plug an ethernet cable into the notebook from the network: the notebook detects the ethernet cable and disconnects the wireless connection to establish a connection to the network over the ethernet cable.

No broadcast storm results.

 

[Marcelmmx]

You would be surprised at the latency having one router on CAT 5 and one on CAT


Yes. Operating systems have default priority levels set to decide which way to access a network in case multiple methods such as wireless and ethernet both exist on a computer/device, and/or multiple network interface cards are present on a computer/device which the operating system can choose from to connect to the network with.

So, how can such a loop come into existence at all? Defect / error?

 

[Ralston18]

Two thoughts at this juncture:

1) Redo your network diagram to show all devices but include each device's name, IP address (DHCP or Static), and subnet mask. If the device's name is personal information then feel free to disguise the name as necessary - as long as you know which device it really is. Indicate if connectivity is wired or wireless.

Also note any reserved IP Static Addresses on the router and verify that the MAC is correct for each static device.

I recognize and agree that a revised diagram will take some time and effort to do. Yet the bigger picture may provide some insight as to what has been happening. Someone may take one look and you will have the "Eureka" moment that is needed to fix the latency problem(s).

2) And per @LWFG001 the notebook probably should switch over. What if the laptop is not doing so? Only one adapter should be enabled. Therefore, my suggestion is to configure the notebook to be either wired or wireless. Then disable the unused adapter. Afterwards, for the time being, simply leave the laptop as is.

Too many variables creeping in and I am still wondering about what happens if the SAT (and now the NAS) are temporarily taken off the network.

Curious about what ping, pathping, and tracert might reveal if you methodically target 2 or 3 network devices. Start without the SAT and NAS, add one or the other back to the network and repeat the tests. Lastly add the remaining device. Note the test results on your network diagram.

 

[Marcelmmx]

Two thoughts at this juncture:

1) Redo your network diagram to show all devices but include each device's name, IP address (DHCP or Static), and subnet mask. If the device's name is personal information then feel free to disguise the name as necessary - as long as you know which device it really is. Indicate if connectivity is wired or wireless.

Also note any reserved IP Static Addresses on the router and verify that the MAC is correct for each static device.

I recognize and agree that a revised diagram will take some time and effort to do. Yet the bigger picture may provide some insight as to what has been happening. Someone may take one look and you will have the "Eureka" moment that is needed to fix the latency problem(s).

2) And per @LWFG001 the notebook probably should switch over. What if the laptop is not doing so? Only one adapter should be enabled. Therefore, my suggestion is to configure the notebook to be either wired or wireless. Then disable the unused adapter. Afterwards, for the time being, simply leave the laptop as is.

Too many variables creeping in and I am still wondering about what happens if the SAT (and now the NAS) are temporarily taken off the network.

Curious about what ping, pathping, and tracert might reveal if you methodically target 2 or 3 network devices. Start without the SAT and NAS, add one or the other back to the network and repeat the tests. Lastly add the remaining device. Note the test results on your network diagram.

Here is the revised network diagram with more details:
Network diagram
Let me know if it's OK like this or if something needs to be added.

 

[bill001g]

This is a very long post now but it I agree with some of the other posters that it almost has to be the switch.

There is no time that is that long normally in a switch. Now if you had a fancy commercial switch that detect things like mac flapping between 2 ports you can get timeouts but I don't remember any time like 8 minutes.

First I will assume reboot the switch takes less than 8 minutes.

Instead of rebooting the switch try unplugging cables. This is more to see if it is the switch itself or if it is because in the process of reboot every port in the switch appears to be unplugged and plugged in and may trigger a action by the device plugged into it.

I would start with the cable that go to the 2 corresponding routers of course only do 1 at a time. If this clears the problem immediately it still does not mean it is not the switch but at least it reduces the devices in your network that could cause the problem.

If it is one of the routers you might want to see if there are any commands you can issue that clear stuff but keep the physical connection up. The DD-WRT one you should be able to clear all the arp and mac tables with commands.

 

[LWFG001]

So, how can such a loop come into existence at all? Defect / error?

The typical problem that creates a broadcast storm is when a switching loop is created because someone plugs both ends of an ethernet cable into the same switch rather than one end to a device and one end to a switch. The link explained it: https://kb.netgear.com/000060475/What-is-a-network-loop

 

[Marcelmmx]

The typical problem that creates a broadcast storm is when a switching loop is created because someone plugs both ends of an ethernet cable into the same switch rather than one end to a device and one end to a switch. The link explained it: https://kb.netgear.com/000060475/What-is-a-network-loop

I meant, how can a single device connect both on WiFi and Ethernet at the same time?

 

[LWFG001]

it almost has to be the switch.

Not necessarily. I've seen mixed cabling produce this problem and he presently IS using a CAT 5e cable on the WiFi B router and CAT 6a on the WiFi A router.

Also, routing that's not set up properly or ip misconfiguration can too.

And he's already taken the steps you're suggesting, trying different ports and re-powering devices, etc...

Putting everything on CAT 6a will remove cabling issues. Testing the switch by swapping it out will determine if it is the switch.

If either of these two fixes solve the problem, great. If not, then I'll dig into IP and routing. But he needs to replace the non-CAT 6a cabling and swap out the switch to test it anyways so let's see if that fixes it first.

P.S. Nice updated diagram Marcelmmx. I'll be using it if replacing the non-CAT 6a cabling with CAT 6a cabling and the switch swap test don't solve the problem.