Slow ethernet iperf3 speeds until reboot

Hi All

Having a bit of an issue with the vero. I installed some cat6 the other day from the media center area to my router to get away from the troublesome wifi extender.

When testing from my pc (in office - ethernet direct to router) through the new cat6 line direct to the vero, I was only only getting 140Mb. WTF is this I cried? And took the vero to the office directly to the router. 940Mb.

OK, back down to the media area. 940Mb. Ok great maybe a patch cable wasn’t quite pushed in properly.

So again this morning I check to find 140Mb/sec. WTF again?
I do 1 thing only, and that is to remotely reboot the Vero, without even touching it. speed check back to 940Mb.
So it seems as if the vero eth is slowing down or dropping out and requiring the occasional reboot?

As with all other requests for support, logs would be more useful when accompanying such a description of an issue.

A log requires a restart, according to the instructions. A restart solves the problem, until the next time it slows down, which may be hours or days, I haven’t been able to reliably reproduce it yet. How would you suggest I log this behavior? Can I somehow produce logs without a reboot?

You can get logs anytime. The only reason we say reboot is to keep the logs short and to the point.

I suggest reboot before you go to bed, leave the device overnight, test to see if it’s slowed up in the morning and grab logs if it has.

Cool, will try that - unsure how long it takes to slowdown yet…

OK, finally after 3 weeks I got the slowdown again. Logs here (but they’re huge):
https://paste.osmc.tv/yizupavoju

The last thing I did this morning at about 10:20 was the iperf3 test, then uploaded.

I’ll check this tomorrow

Cheers

Sam

Hi Sam - any clues in the logs?

Jun 29 23:31:27 osmc nmbd[646]:   query_name_response: Multiple (2) responses received for a query on subnet 192.168.1.24 for name WORKGROUP<1d>.
Jun 29 23:31:27 osmc nmbd[646]:   This response was from IP 192.168.1.1, reporting an IP address of 192.168.1.1.
Jun 29 23:32:44 osmc sshd[4293]: pam_unix(sshd:session): session closed for user osmc

alot of this before shutdown seems there are some conflicts in your network that needs your attention.

@Toast - can you elaborate? the vero is .24 and the gateway is .1 Not many devices on the system which is why I start all fixed IP’s at 20 and dhcp takes place from .3 upwards.
And why does the log refer to “subnet” .24? that’s weird. The vero’s ip is .24

Is this snippet repeated many times? what would the error indicate?

Hi @floatingkiwi — sorry this slipped through the cracks

Can you check the device on DHCP as a quick test?

Sam

Yes I’ll change to DHCP and keep testing.

1 Like

OK, I managed to get this ethernet slowdown to happen within a day. Again, on a GB network, restart the vero, perform iperf3, perfect 938MB tests.
Wait a few hours, repeat ad nauseum until eventually the vero ethernet speed drops to something around 100-200MB.

The ONLY thing that fixes this is a vero reboot. All other devices have been rebooted during all these experiments, cables swapped etc, routers, switches.

Logs:
https://paste.osmc.tv/ahafenipov

This jumps out at me in your logs:

2020-02-26 12:50:47.953 T:3695096544   DEBUG: CAddonSettings[plugin.program.iptv.merge]: loading setting definitions
2020-02-26 12:50:47.953 T:3695096544   DEBUG: CAddonSettings[plugin.program.iptv.merge]: trying to load setting definitions from old format...
2020-02-26 12:50:47.955 T:3695096544   DEBUG: CAddonSettings[plugin.program.iptv.merge]: loading setting values

Those 3 lines appear over 10,000 times in your log. I’d suggest testing with that addon disabled.

Also, this happens quite a bit, about 45 minutes:

Feb 26 12:55:35 osmc nmbd[782]: [2020/02/26 12:55:35.572612,  0] ../source3/nmbd/nmbd_namequery.c:109(query_name_response)
Feb 26 12:55:35 osmc nmbd[782]:   query_name_response: Multiple (2) responses received for a query on subnet 192.168.1.24 for name WORKGROUP<1d>.
Feb 26 12:55:35 osmc nmbd[782]:   This response was from IP 192.168.1.1, reporting an IP address of 192.168.1.1.

And this is odd:

Feb 26 13:00:35 osmc nmbd[782]:   This response was from IP 192.168.1.1, reporting an IP address of 192.168.1.1.
Feb 26 13:01:30 osmc ntpd[1011]: 193.27.208.100 local addr 192.168.1.24 -> <null>
Feb 26 13:03:23 osmc ntpd[1011]: 90.187.7.5 local addr 192.168.1.24 -> <null>
Feb 26 13:05:32 osmc ntpd[1011]: 37.235.209.151 local addr 192.168.1.24 -> <null>

Your logs aren’t showing anything obvious.

Next time it happens, can you disable Ethernet in My OSMC -> Networking? Wait half a minute or so, re-enable? Does it resolve the issue?

Sam

I will try sam - sometimes it takes days to slow down, this time was real fast within the day.

192.168.1.1 is my router.

Hey guys,
I just saw this thread and thought about posting my experience as well.
Basically the same what happens to floatingkiwi’s Vero, does happen to my Vero.
It seems like whenever the device is idling for a while the connection degrades to a 100 Mbit/s connection. (The user interface of my router displays "LAN 1 with 100 Mbit/s)

When it happens the next time, I’ll try what @sam_nazarko suggested.

Will need to see logs to advise.

Could you guys post the output of: ifconfig
If a device downgrades the connection speed, there are usually 2 reasons for that:

  1. The remote end requested it (Switch goes to green mode or so)
  2. The traffic quality is bad (overruns, drops, overruns)

One can see these. One my vero I have some dropped packets, but nothing alarming:

eth0: flags=-28605<UP,BROADCAST,RUNNING,MULTICAST,DYNAMIC>  mtu 1500
    inet 10.0.3.242  netmask 255.255.254.0  broadcast 10.0.3.255
    ether c4:4e:ac:28:57:97  txqueuelen 1000  (Ethernet)
    RX packets 221073  bytes 139075303 (132.6 MiB)
    RX errors 0  dropped 3782  overruns 0  frame 0
    TX packets 105557  bytes 15478661 (14.7 MiB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
    device interrupt 40 

You see here that I have some dropped packets.

I just restarted the device, did ifconfig and it shows indeed some drops

eth0: flags=-28605<UP,BROADCAST,RUNNING,MULTICAST,DYNAMIC>  mtu 1500
    inet 192.168.178.25  netmask 255.255.255.0  broadcast 192.168.178.255
    ether 90:0e:b3:01:c7:31  txqueuelen 1000  (Ethernet)
    RX packets 527  bytes 42394 (41.4 KiB)
    RX errors 0  dropped 296  overruns 0  frame 0
    TX packets 241  bytes 32249 (31.4 KiB)
    TX errors 0  dropped 0 overruns 0  carrier 0  collisions 0
    device interrupt 40