RB2011UAS-RM - auto negotiation problem

Other topics MikroTik hardware questions
1

Hi guys,

I have just installed a RB2011UAS-RM for a customer and discovered a problem with the LAN speed negotiation.

I connected LAN port 3 on the RB to a BT FTCC modem (ECI V-2FUb/I Rev.B) and discovered it negotiated at 10 Mbps. First thought was of course bad cable. Having replaced the cable the same problem exists. I started fiddling around with switching the auto negotiation off and full duplex. Forcing 100 or 1000 Mbps causes the link to be lost.

I then moved over to port 6 on the RB - same problem with auto negotiate. I think I may have been able to force 100 Mbps but only half duplex (struggling to remember exactly what I tried).

Next I tested the BT FTTC modem connected to a laptop using the same cable - full 1000 Mbps
Then I connected the RB to the laptop using the same cable - full 1000 Mbps on port 3 and 100 Mbps on port 6.

Is this some kind of bug in the RB hardware or incompatibility between devices?

The BT modem is locked and as such not a single setting can be changed on that device.

I’ve had to put a standard 10/100 switch between the two devices just to get above 10 Mbps.

Tried ROS 5.24 and v6.0rc11 with no change.

Any suggestions?

Cheers

Henry

Update 24/02/2013: Found another Mikrotik user from the UK with the same problem and similar equipment. The thread is here (with no positive outcome)

Someone with a similar port speed issue here (with no positive outcome)

And another thread with port speed problems here (with no positive outcome)

And finally some mention of it here here

2

I wouldn’t be surprised if Mikrotik have not been able to reproduce the problem reliably. There have been problems of the other OpenReach modem the HG612 having auto-neg problems on some PC ethernet card drivers although I have not seen such comments regarding the ECI V-2FUb/I Rev.B. The other 3 links don’t really help a great deal - rather sketchy details.

I had the chance to look at a 2011 which had 4 ATAs attached and two had come up at 10 Mb. I thought maybe that situation would have provided some hard information but in fact it turned out that two of these brand new ATA were set to 10Mb whereas the other two were set to 100Mb and the installer had simply not considered that possibility.

It would be useful to know if an unlocked ECI V-2FUb/I Rev.B reveals any settings which can be adjusted to avoid the problem.

3

I wouldn’t be surprised if Mikrotik have not been able to reproduce the problem reliably

Perhaps I could loan some hardware to Mikrotik so they can investigate further? I’m sure this problem is going to affect everyone in the UK with an FTTC modem.

It would be useful to know if an unlocked ECI V-2FUb/I Rev.B

I had a look at the screenshots of the web interface, there aren’t any settings relating to port speed.

I’m going to try a crossover cable today as an extreme long shot.

Also, for others thinking of going down this path, I’m having stability issues with the pppoe connection. Every 24hrs or so the connection will drop and the Mikrotik is unable to re-establish it. From the logs I can see the broadcast packet going out but no response. Restarting the BT modem solves the problem.

It would seem the BT modem and the Mikrotik router just don’t get along. Unfortunately this is the solution we have invested in! :open_mouth:

4

We have worked around this problem with our customers by them using port 10 on the RB2011 set at autonegotiate they get 100 Mbps full duplex to the BT supplied FTTC modems.

We have reported this to MikroTik support, and supplied supouts etc using both common FTTC modems as supplied by BT.

Nick.

5

nickshore

Is it possible to get ticket number of the problem?

6

Thanks for the info Nick. We tried port 10 this afternoon and in our case it doesn’t help at all.

Currently we have 100 Mbps half duplex with auto neg off on port 6

7

I too have problems with the auto-negotiation with the RB2011.

For some reason my ESX server directly connected to it (Intel PRO/1000 PT) often comes up as 100Mbit instead of 1Gbit. What I do is connect using winbox, tick off auto negotiation on that port for 5 seconds and then click it on again, it will immediately take up the right speed of 1Gbit and work. As long as I don’t reboot the RB2011, it stays correct.

The same sometimes happens with my UBEE cable modem. It should sync at 1Gbps but sometimes comes up as 100Mbit. The same trick (tick auto neg. off and after 5 seconds on again) fixes the problem till I have to reboot.

I’ve have this through Version 5.18 up to 6.0rc11 right now. No clue why it happens. Different cables or anything don’t help anything. But with the little trick, it’s stable and speeds are good. I do really get 1Gbit (117MB/sec using Iperf) so the switch chip is working correctly, it just the sync after booting that doesn’t work correctly.

8

Mikrotik have informed me they are aware of the problem and investigating a fix, I will of course let you know…

9

[Ticket#2013012366000781] RB2011 ethernet port autosync problems
I reported same problems but the last correspondence was due from to me to MT and I’ve not had time to reply. Same equipment as other have reported here. BT VDSL modems both the ECI and Huawei types

10

Interesting to know it’s ECI and Huawei

I’m still running at 100 Mbps half duplex at the moment but even that doesn’t seem totally stable. Here’s what happened to the FTTC connection last night out of nowhere (having been up since 17:30)

2013.02.27-02:01:41 <172.16.1.252>: route,debug,event debug: Interface change
2013.02.27-02:01:41 <172.16.1.252>: route,debug,event debug: interface=ether6 - WAN2
2013.02.27-02:01:41 <172.16.1.252>: route,debug,event debug: status=UP
2013.02.27-02:01:41 <172.16.1.252>: route,debug,event debug: mtu=1500
2013.02.27-02:01:41 <172.16.1.252>: pppoe,debug pppoe: ether6 - WAN2: terminating sessions: interface state changed

There was nothing in the logs saying the interface went down, which seems odd to me. It went ‘up’ at 17:30 and then ‘up’ again at 2:01 with no ‘down’ in between. :confused:

11

[Ticket#2013012366000781] RB2011 ethernet port autosync problems

Bazzer
Am I correct in thinking from the number that your ticket was raised on the 23rd Jan 2013?

12

Any news from Mikrotik about when this will be fixed? Makes it difficult for us to recommend this unit with the problem still outstanding. :confused:

Thanks

Henry

13

We have ordered equipment for additional testing. It should arrive soon, then we could comment more about this problem.

14

Hi guys,

Sorry to keep on at you… but any progress at all? Really happy with the RB2011 but this one issue is letting it down!

Thanks,

Henry

15

Dont know if this is helpful to you but my RB450G is connected to the BT supplied VDSL modem and auto negotiates at 100mb full duplex on port one.

This has worked fine for me from day one and i get my full download/upload bandwidth.

16

AAAH! I’m having this problem too! I thought it was my equipment!

It depends on what hardware I am connecting to the port. Cheap equipment (Dlink routers, Chinese cameras, etc) all need to be manually set to 10meg half duplex to connect. This also makes a difference depending what port group I use. If I use the 100meg ports, the problem is more common. The gigabit ports do the same problem, but it seems not as often.

17

Any progress on this?

We are seeing it more and more commonly.
Again ONLY on 2011 boards.

18

I have same type of issue against MT own equipment (http://forum.mikrotik.com/t/problem-of-gigabit-port-between-cisco-and-mikrotik/59912/20).
Note: According to the Wikipedia sources the autonegotiation is a must for the GBit/s LAN (http://en.wikipedia.org/wiki/Gigabit_ethernet#1000BASE-T) and disabling it is illegal.

19

Hi,

I have a RouterBoard 2011UAS-RM which I am using with an Openreach FTTC modem (HUAWEI Echolife HG612).

I am seeing 100MB Auto Negotiation on Ether 1, the RB is running 3.04 Firmware and RouterOS 5.24.

20

Have the same problem with RB600, RB411 and also X86 computer with RouterOS. And also many different HW variants on the other end from enterprise cisco switch to few SOHO switches do the same. Even if I turned off autonegotiation at all on both ends of the wire RouterOS was reporting wrong speed on status page after some time and ethernet link had about 80% packet loss. Turning ON and again OFF the autonegotiation solves the problem for unpredictible amount of time. From that i guess it is the SW problem we are now using version 4.10 so I will try to upgrade and let we see. What versions are you experiencing this with?