49 replies to this topic

[bleskone]

thanks guys for your comments,

my next findings are, without any usb key plugged

 

I have clear machine (without disconnects in log), usually after deep standby and wake up cycle,

 

if I soft reboot box - disconnects are again in the log

if I power off/on box with button - disconnects are again in the log

the only case when I can get rid of them is: from any state - going to deep standby, wait some seconds, and wake up with remote, - in this procedure - disconnects does not appear (also log is clear) ....

[Meega]

thanks guys for your comments,

my next findings are, without any usb key plugged

 

I have clear machine (without disconnects in log), usually after deep standby and wake up cycle,

 

if I soft reboot box - disconnects are again in the log

if I power off/on box with button - disconnects are again in the log

the only case when I can get rid of them is: from any state - going to deep standby, wait some seconds, and wake up with remote, - in this procedure - disconnects does not appear (also log is clear) ....

I can confirm this ,even with usbhub ,usbstick, usbhdd and ledlights connected.

(after testing a couple of day's)

 

ps ,i don't have any problems with it ,except when i want to view the dmesg log. 

[bleskone]

thanks for sharing experiences..

 

my next findings and probably the last ones with this

 

if I have USB key plugged in rear port - disconnects (messages in log) happen all the time

if I have USB key plugged in front port - I can get rid of them

 

I made my personal rule, after all actions: soft reboot, hard reboot, power outages, - I always go to deep standby and wake up

I use box heavy with NAS - parallel recordings to and watching from network device, in this situation some disconnects made quite long timeouts 30 sec -1 min, which was too annoying

 

 

I am sorry to say, but with informations that this occasionally happen on majority "new" Xtrend models ET8000, ET10000, ET8500,.. and the problem is left out untouched, this is show stopper for me and probably next Xtrend's are not for me anymore ... at least current models, but the firm philosophy always sending signals to users ...

[Erik Slagter]

You can now attach a gbit USB ethernet adapter.

[bleskone]

You can now attach a gbit USB ethernet adapter.

 

ok, thanks, are there any recommended usb-->network adapers working in Openpli, I know - usb2 limit is 480 Mbit - so there will be only cca half usefull bandwith but still enough

[Erik Slagter]

I have one in use myself. The througput isn't amazing, but it's not bad as well, expect about 300-400 mbps. Always 2-5x better than the 100 mbps interface.

 

The adapter should have an ASIX chip for it to be supported. The good news is, it seems almost all of them do (for gbps adapters). I bought myself one when I was finally completely done with the gbic nic of my et8500 constantly flapping.

[bleskone]

thanks for info, assuming that this one should be OK .... GigaBlue USB GigaBit Adapter

[jpuigs]

My ET10000 box never had this problem, but since a few days, in dmesg output I  see it too:

 

bcmgenet bcmgenet.1 eth1: link down
bcmgenet bcmgenet.1 eth1: link up, 1000 Mbps, full duplex
bcmgenet bcmgenet.1 eth1: link down
bcmgenet bcmgenet.1 eth1: link up, 1000 Mbps, full duplex
bcmgenet bcmgenet.1 eth1: link down
bcmgenet bcmgenet.1 eth1: link up, 1000 Mbps, full duplex
bcmgenet bcmgenet.1 eth1: link down
bcmgenet bcmgenet.1 eth1: link up, 1000 Mbps, full duplex
bcmgenet bcmgenet.1 eth1: link down
bcmgenet bcmgenet.1 eth1: link up, 1000 Mbps, full duplex

 

Has anybody discovered anything more ?

[WanWizard]

Your box is up to date? Because there has been a driver update a few days ago with changes to the bcmgenet code...

[jpuigs]

Last updated 18-04-2016 14:04. it was updated yesterday, and they're still appearing....

[WanWizard]

So it might be related to the last driver update. Are other people having the same issue?

[socks]

In my case this is related to heat problems.

I have to switch off the box, wait for about 10 minutes to let it cool down, and then switch it on again. Then the flapping is gone.

When I just reboot the box (soft reboot or switch off/on) without letting it cool down the GigabitEthernet VLAN interface is flapping continuously.

[jpuigs]

It seems that after updating box two days ago, problem has dissapeared.

I have seen no more disconnection/reconections again.

Had driver been updated again ?

[betacentauri]

No, not yet.

[jpuigs]

It's really strange.

I've never had this problem, since some days ago, I think it was aprox. when kernel was updated from 3.XX to 4.

And now, without updating (driver) it dissapears itself......

I haven't installed or unistalled anything, except for online updates. So it has to be something related to any file.

Otherwise we'll have to believe in miracles.

Edited by jpuigs, 23 April 2016 - 23:41.

[socks]

jpuigs: I also thought about many reasons before I found out that it is heat related.

If you have interface flapping next time please try to switch off the box and let it cool down for 10 minutes. This way the problem is gone until the next reboot.

Btw, what temperature is shown on your box under Menu > Information > About > CPU?

[jpuigs]

66 º

Edited by jpuigs, 24 April 2016 - 18:57.

[anudanan]

Are the driver for USB to ethernet adapter with ASIX chip  inside the image?

If not, how can I install it later?

[WanWizard]

Afaik they aren't. And you need to compile them when you compile the image, which means you need to add an OpenEmbedded recipe for it.

 

If you do an "lsusb", what is the USB id of the adapter?

[betacentauri]

Some drivers are included in the kernel and some modules are on the feed:
https://github.com/X...defconfig#L1222