Vero 4K + network speed major disappointment

Forcing full duplex with auto negotiation disabled will cause the switch at the the other end of the link to assume half duplex, this will cause a duplex mismatch causing extremely poor performance in one direction.

Repeat the test with half duplex and it should be fine in both directions:

ethtool -s eth0 speed 100 duplex half autoneg off

Never force full duplex with auto-negotiation off at one end of the link without also simultaneously forcing full duplex at the remote end. If you don’t have control of the configuration of the other end of the link, always use half duplex when disabling auto-negotiation.

1 Like

Means next to nothing to me, I’m afraid. I just typed in what the dude above asked, in the hope of fixing it. Turns out I’m not alone, and there’s a number of people with the same issue.

If you fancy telling me what it should be set to, so I can ensure it is back to normal, that would be helpful.

I’ve spent more time messing with this network stuff than I did building my previous box from scratch.

yup gotta agree with @DBMandrake (is that an old distro or wizard reference?) that forcing 1000mb full duplex rarely works on linux - better to let it autoneg or you’ll end up with 10/half

1 Like

My issue is that the Vero 4K+ is not negotiating a gigabit connection to my switch. I have tried different ethernet cables and two different gigabit switches:

  • Ubiquiti UniFi US-8-150W (managed, port link negotiation set to auto)
  • HP ProCurve 1400-8G (unmanaged)

I also tested the kernel from ([TESTING] Vero 4K / 4K + video improvements), no changes.

Here are the logs:
https://paste.osmc.tv/mojepoxoqe

Here’s the output of ethtool (with UniFi switch):

osmc@osmc:~$ sudo ethtool eth0
Settings for eth0:
Supported ports: [ TP MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supported pause frame use: Symmetric Receive-only
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised pause frame use: Symmetric Receive-only
Advertised auto-negotiation: Yes
Link partner advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
Link partner advertised pause frame use: No
Link partner advertised auto-negotiation: Yes
Speed: 100Mb/s
Duplex: Full
Port: MII
PHYAD: 0
Transceiver: external
Auto-negotiation: on
Supports Wake-on: ug
Wake-on: d
Current message level: 0x0000003d (61)
drv link timer ifdown ifup
Link detected: yes

Any suggestions?

Are both times the link partner speeds reported as follows?
Link partner advertised link modes: 10baseT/Half 10baseT/Full 100baseT/Half 100baseT/Full

Both switches gave:
ā€œLink partner advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Fullā€

Here’s the ethtool output for the HP switch (see above for UniFi):

osmc@osmc:~$ sudo ethtool eth0
Settings for eth0:
Supported ports: [ TP MII ]
Supported link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Supported pause frame use: Symmetric Receive-only
Supports auto-negotiation: Yes
Advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
1000baseT/Half 1000baseT/Full
Advertised pause frame use: Symmetric Receive-only
Advertised auto-negotiation: Yes
Link partner advertised link modes: 10baseT/Half 10baseT/Full
100baseT/Half 100baseT/Full
Link partner advertised pause frame use: Symmetric
Link partner advertised auto-negotiation: Yes
Speed: 100Mb/s
Duplex: Full
Port: MII
PHYAD: 0
Transceiver: external
Auto-negotiation: on
Supports Wake-on: ug
Wake-on: d
Current message level: 0x0000003d (61)
drv link timer ifdown ifup
Link detected: yes

That’s really odd. I assume you have validated the cable and the port reports correctly if you plug in another Gigabit device, or?

Maybe check for errors with ethtool -S eth0
Also maybe try to set speed ethtool -s eth0 speed 1000 duplex full

Withethtool -S eth0 | grep -i err everything is 0, full log at: https://paste.osmc.tv/efemeqegoq

I set the UniFi switch port 4 to 1000/full duplex and did the same for vero 4k+ with ethtool -s eth0 speed 1000 duplex full. Vero 4k+ fails to get any link to both the HP switch (which is unmanaged, cannot configure it) and the UniFi switch. UniFi switch port 4 ↔ HP switch gives a gigabit link.

I assume same cable being used?
Than I am out of ideas

Yes, it was the same cable. I’m also out of ideas, I’ve tried three different cables and now three different switches (last one was Asus RT-AC66U router/switch, same problem).

Maybe @sam_nazarko has some magic ideas

Hi @rusakko

Sorry to hear you are experiencing a problem here.
Do you have another device / environment you can test with so we can see if the issue is localised?

I can produce a kernel with some debug messages so we can see why the negotiation is failing.

Sam

Hi,

Which device are you referring to? The Vero 4k+? I only have access to this unit.

Sure, I can test a kernel with debug messages.

With regards to the network environment, I tested three different switches and cables so it’s probably not that.

I have an Asus RT-AC68U in AP mode. I know yours is an AC66U, but still… If I get a moment today I’ll set up my second Vero 4K+ directly connected to that, and report my iperf results if that’ll help?

For me the issue is that I don’t even have a gigabit ethernet connection between my switch and the vero 4k+. Link speed is 100Mbit/s so that’s the fastest speed I could get, haven’t really done any performance tests before my link speed is 1Gbit/s.

I suspect there could be something wrong with my Vero 4K+ unit.

Hi @sam_nazarko,

I’ve tested three different gigabit switches and cables. Seems unlikely that it’s related to either of those, different switch manufacturers and bought years apart from each other (HP is 10 years old, Asus router is 5 years old and the Ubiquiti Unifi is brand new). I don’t think I’ve ever had this kind of problem with a gigabit ethernet device before, usually they just work ā„¢.

I don’t have access to another Vero 4K+ so I can’t rule out that this unit is faulty.

What’s the next step, the test kernel you mentioned?

Hi,

If you tested with 3 switches and cables, I cannot conceive it being an environmental issue.

Please email support@osmc.tv or give me your order # and I will raise a ticket on your behalf and we will replace this unit for you promptly.

I apologise for the inconvenience.

Sam

Hi,

Sounds good, my order number is #18949.

I’ll raise a ticket on your behalf in a couple of hours.