Ethernet port speed dropping from 1Gbps to 100Mbps (AC5300 & 384.12)

[norp90]

Hi,

I running a AC5300 & 384.12. Has anyone else experienced the Ethernet ports dropping from an initial 1Gbps port speed when the cable is first plugged in to 100Mbps and staying there? This is the speed as reported in both the AC5300 GUI and the upstream switch it is connected to. I'm not seeing anything obvious in the AC5300 logs. If I unplug and plug the Ethernet cable back in the port will return to 1Gbps again. I don't believe it is a cable issue.

Thanks.

 

[SheikhSheikha]

Hi,

I running a AC5300 & 384.12. Has anyone else experienced the Ethernet ports dropping from an initial 1Gbps port speed when the cable is first plugged in to 100Mbps and staying there? This is the speed as reported in both the AC5300 GUI and the upstream switch it is connected to. I'm not seeing anything obvious in the AC5300 logs. If I unplug and plug the Ethernet cable back in the port will return to 1Gbps again. I don't believe it is a cable issue.

Thanks.

I have two AC5300 running - never seen them behave like you describe. Why not update to 384.12 and do a full reset (M&M, erase NVRAM etc). Okay, it is maybe some work but in all cases worthwile to exclude any other reason than a hardware failure (either the port or the cable) ...

 

[ApexRon]

I hope you have noticed that from the GUI on the AC5300 you cannot set the port speed. This is because the port speed is negotiated, though a cable issue could also force a lower port speed.

"... speed as reported in both the AC5300 GUI and the upstream switch ..."​

As a first step, I would highly recommend trying a known working 1Gbps cable. Also be sure that there are no dust bunnies in either port and the port and plug connectors don't have any pins that are pushed back or bent.

Since you state that the speed is 1Gbps when you plug the cable in but then goes down to 100Mbps, I suspect the Auto-Negotiation handshaking process is either detecting a cable capabilities issue or there is some sort of Auto-Negotiation incompatibility between the switch and the AC5300, assuming the switch sets port speed that way.

For me, prior to 384.13, I noticed that my port speed would toggle between 1Gbps and 100Mbps when my AppleTV was in standby. Moving to 384.13 has resolved.

 

[norp90]

Still having this issue. I've changed the cable for a brand new cat6 patch cable between router and switch and upgraded the router to 384.13, no change, port speed as shown in the router homepage status section is dropping to 100Mbps and I don't know why or if the speed it really dropping or just UI reporting incorrectly. If I unplug the cable and plug in again, refresh the webpage the speed will be back up to 1Gbps.

Edit: The port speed is not dropping back to 100Mbps straight away so I don;t suspect the auto-negotiation process is the cause here. I haven't been able to tell when exactly it happens/what triggers it but it will stay at 1Gbps for some period of time and later drop back.

 

[ColinTaylor]

It sounds like it is caused by the "upstream switch it is connected to". What is this device? Maybe that is rebooting itself or has some sort of green/power-saving mode that is incompatible with the router.

 

[ApexRon]

The port speed is not dropping back to 100Mbps straight away so I don;t suspect the auto-negotiation process is the cause here. I haven't been able to tell when exactly it happens/what triggers it but it will stay at 1Gbps for some period of time and later drop back.

Welcome back from your sabbatical ...
Auto-Negotiation is a continuous process. Please refer to these two articles:

https://en.wikipedia.org/wiki/Autonegotiation
https://www.cisco.com/c/en/us/support/docs/lan-switching/ethernet/10561-3.html​

There are many more sources of information but the point is that once you truly understand how Auto-Negotiation works troubleshooting your issue may be easier.

Additionally, as @ColinTaylor suggested, now that you have replaced the cable. Knowing what is connected to the other end of the cable will help us help you.

 

[norp90]

The upstream switch is a Ubiquiti EdgeSwitch 24 250W https://dl.ubnt.com/guides/edgemax/EdgeSwitch_ES-24_QSG.pdf and the port is set to Auto-negotiate.

 

[norp90]

It sounds like it is caused by the "upstream switch it is connected to". What is this device? Maybe that is rebooting itself or has some sort of green/power-saving mode that is incompatible with the router.

Thanks, out of interest what makes you lean towards it being an issue with the switch and not a bug in the firmware of the router? ApexRon mentioned above he was experiencing something similar prior to 384.13. I agree, could be on issue on either end. I'm not having this issue with other devices connected to the switch though.

 

[ColinTaylor]

Thanks, out of interest what makes you lean towards it being an issue with the switch and not a bug in the firmware of the router?

I'm only thinking that because a) you've replaced the cable which is the most common cause, and b) you've updated the firmware which the others said fixed their problem. There's also been a few reports (I can't remember for which router models) where the router simply didn't play nice with another device and inserting an intermediate switch fixed the problem. It's not unheard of to have interoperability issues between different manufactures' chipsets.

 

[norp90]

I'm only thinking that because a) you've replaced the cable which is the most common cause, and b) you've updated the firmware which the others said fixed their problem. There's also been a few reports (I can't remember for which router models) where the router simply didn't play nice with another device and inserting an intermediate switch fixed the problem. It's not unheard of to have interoperability issues between different manufactures' chipsets.

Interesting, thanks.

 

[ApexRon]

I concur with @ColinTaylor upgrade firmware on Asus. However, a simple solution may be just to make that port on the switch 1Gbps. Here is the CLI format for doing that on your switch:

speed
Use this command to enable or disable auto-negotiation and set the speed that will be advertised by that port . The half-duplex and full-duplex parameters allow you to set the advertised speed for half-duplex and full-duplex modes .
Use the auto keyword to enable auto-negotiation on the port . Use the command without the auto keyword to ensure auto-negotiation is disabled and to set the port speed and mode according to the command values . If auto-negotiation is disabled, the speed and duplex mode must be set .

Default

Auto-negotiation is enabled.​

Format

speed {auto {40G|10G|1000|100|10} [40G|10G|1000|100|10] [half-duplex | full-duplex] | {40G | 10G | 1000 | 100 | 10} {half-duplex | full-duplex}}​

Mode

Interface Config​

Here is the manual I got it from; https://dl.ubnt.com/guides/edgemax/EdgeSwitch_CLI_Command_Reference_UG.pdf

Also, the current firmware for your switch is v1.9.0 . Consider making that current as well. https://www.ui.com/download/edgemax/edgeswitch

 

[baz4096]

Just made an account specifically to say that I too have this issue. I've tried multiple cables, and a new switch (mine is unmanaged) and still have the same problem. I've had two months of 1Gbps, then today, it's reporting as 100Mbps between my ASUS RT-AC5300 and my nearly brand new Netgear GS348. This is very frustating as my internet connection is 370Mbps, which is now artificially cut down to around 80.
I have a firmware update to try (although I'm only one minor version behind), and whilst I've power cycled both router and switch, I haven't powered down both at the same time, which I will try when the wife isn't in a Zoom meeting.

 

[ApexRon]

I've had two months of 1Gbps, then today, it's reporting as 100Mbps between my ASUS RT-AC5300 and my nearly brand new Netgear GS348.

The negotiated speed between devices is determined by the physical attributes of the cable and auto-negotiation.

Autonegotiation can be used by devices that are capable of more than one transmission rate, different duplex modes (half duplex and full duplex), and different standards at the same speed (though in practice only one standard at each speed is widely supported). Each device declares its technology abilities, that is, its possible modes of operation, and the best mode is chosen from those shared between them, with higher speed preferred over lower, and full duplex preferred over half duplex at the same speed.

Parallel detection is used when a device that is capable of autonegotiation is connected to one that is not. This happens if a device does not support autonegotiation or autonegotiation is administratively disabled on a device. In this condition, the device that is capable of autonegotiation can determine and match speed with the other device. This procedure cannot determine duplex capability, so half duplex is always assumed.

Other than speed and duplex mode, autonegotiation is used to communicate the port type (single port or multiport) and the master-slave parameters (whether it is manually configured or not, whether the device is master or slave if this is the case, and the master-slave seed bit otherwise).

From: https://en.wikipedia.org/wiki/Autonegotiation

Based on the description you provided, it would appear to be a:

1. Cable issue - Purchase a quality Cat 6 cable, making sure that the cable is labeled as such.
2. Noise - Electrical noise is a possibility, especially if your not using the latest Cat 6 cable
3. Auto-negotiation incompatibility between ASUS and Netgear. Unfortunately, a firmware upgrade may not resolve if the incompatibility is at the chipset level. None the less, upgrade to latest firmware.

 

[baz4096]

Hi ApexRon

Thanks for the reply. Currently I have good quality sockets and patch bay, with CAT6 throughout, including both patch cables, but I suffer the same issue if I replace the entire run with a single new CAT6 lead. My switch is unmanaged, so no firmware there. I’ll update my router again to see if that helps, but the patch notes only show security updates. I’ll also try a different socket on my AC5300 to see if maybe there’s a faulty socket there. If it continues, the only choice I have will be to replace the router itself, but I’m loath to do that as it’s not that old, and wasn’t that cheap Everything else has been replaced. Not sure how long this has been going on for as until recently I was on a 60Mbps internet connection and hadn’t noticed any significant drop in speed. Now I’m on 370 it’s obvious when it happens. Also, whilst everything is CAT6, CAT5e should be able to support 1Gbps, or am I wrong in thinking that?

 

[ApexRon]

... Not sure how long this has been going on for as until recently I was on a 60Mbps internet connection and hadn’t noticed any significant drop in speed. Now I’m on 370 it’s obvious when it happens. Also, whilst everything is CAT6, CAT5e should be able to support 1Gbps, or am I wrong in thinking that?

CAT 5 is more susceptible to crosstalk which could cause a drop in speed.

 

[baz4096]

My sockets are wired with solid core copper CAT6 (with the + shaped insulator separating the pairs). Have done the firmware update on my router, but that’s made no difference. Tomorrow I’ll power everything off for a few minutes. I’m wondering if the auto negotiation isn’t retrying fully if only one side gets power cycled at a time. Grasping at straws now...

 

[ApexRon]

I assume you have tried different ports on the ASUS and Netgear. A troubleshooting technique I would try Is to minimize what’s connected to each and then, after a period of time, connect one cable and wait again. Keep repeating until all returned or one re-connection causes a change of speed.

 

[baz4096]

OK, so powering down both devices for a few minutes helped me here. When they were switched back on, the connection between them was back at 1Gbps. I wonder if there was some additional interference or power spike that caused them to renegotiate a slower speed, as it's happened on both my built-in sockets/patch panel and a separate, new CAT6 cable.
The last time this happened to me was a couple of months ago, so to say it's intermittent would be an understatement, so in the meantime I've stuck a spare Gbps5 port switch next to my router, with a short patch cable between them. Maybe I'll have better luck with this setup. As both my 48 port and 5 port switches are unmanaged, I'm now reliant on spotting an orange LED vs green as I walk past it...

On my ASUS AC-RT5300, I've created a script that hooks into the call used by the web config pages to display the port status, which I've placed in the jffs folder (I think this persists between reboots), and call remotely via a cron from a second linux box. If the output matches what's expected, nothing happens, otherwise I get an email.

 

[ApexRon]

OK, so powering down both devices for a few minutes helped me here. When they were switched back on, the connection between them was back at 1Gbps. I wonder if there was some additional interference or power spike that caused them to renegotiate a slower speed, as it's happened on both my built-in sockets/patch panel and a separate, new CAT6 cable.
The last time this happened to me was a couple of months ago, so to say it's intermittent would be an understatement, so in the meantime I've stuck a spare Gbps5 port switch next to my router, with a short patch cable between them. Maybe I'll have better luck with this setup. As both my 48 port and 5 port switches are unmanaged, I'm now reliant on spotting an orange LED vs green as I walk past it...

On my ASUS AC-RT5300, I've created a script that hooks into the call used by the web config pages to display the port status, which I've placed in the jffs folder (I think this persists between reboots), and call remotely via a cron from a second linux box. If the output matches what's expected, nothing happens, otherwise I get an email.

Good to hear.

I just had an interesting experience. My ASUS router, RaspberryPi (Ethernet only), and Apple router are all within four feet of each other, no LAN issues and all were at 1 Gbps. Then I installed a WYZE base station and my RaspberryPi connection went down to 100 Mbps, rebooting did not resolve. I have heard of 2.4 GHz injecting noise into Ethernet connections, so I installed a CAT 7 cable in place of my CAT 6. Problem solved.

 

[baz4096]

I might have to give that a try at least for now I have a bunch more sockets on my hub shelf.