Community discussions

 
nissandata
newbie
Topic Author
Posts: 33
Joined: Fri Dec 03, 2010 7:20 pm

RB2011UAS stopped working during heavy traffic on SFP

Sat Dec 29, 2012 1:34 pm

Hi,
I have quite limited info on this problem ATM, so taking a long shot here to see if someone could have any info on the matter.

We have just set up two RB2011UAS-IN which we basically use as fibre converters/switches.

I have a mm-fiber between them and then via VLAN whe practicly have mapped eth1 on the first RB2011 to eth1 on the second RB2011. Same with eth2, eth3 etc.
Through this link one costumer receives their internet connection, and another costumer (on the other side of the road) has a NAS for their backup-environment.

This was installed a few days ago, and we have done some test-backups pushing the link to its full 1Gbit speed. Everything has worked fine. But yesterday we ran two simultaneous backups through the link. And after about 40 min the connection between the RB2011's stopped working.

Due to holidays and limited physical access to the devices I was not able to connect to any of the RB's to troubleshoot or see the logs. But the costumer restarted one of the RB's and everything went up again.

- Is it possible that the RB's shutsdown the SFP-port if alot of packet loss or collisions/errors occurs?
- Have you seen any known bug which could be related to this? I think v5.22 is installed on the devices, the devices were delivered two weeks ago.

Some additional info
- The fiber has been in use for about 2 years without problems, but ran from an HP ProCurve to a fiber converter before we installed the RB's. But at the time it was only used to deliver 30mbit internet to the costumer (even though the fiberlink was 1Gbit)
- The SFP-modules used are 3rd-party cisco-modules, the reseller for routerboard in Sweden recommended this very SFP since it was what they use them self.
- This has just happend once, and it is possible that this is just an coincidence, that it went down at the same time as the high load occurred, but I think it's a reasonable guess that it is connected.

- The only configuration on the devices are:
5 bridges,
5 vlans,
vlans connected to SFP, one bridge and one eth-interface each
1Gbit on SFP whith auto-negotionation disabled
identity name
 
nissandata
newbie
Topic Author
Posts: 33
Joined: Fri Dec 03, 2010 7:20 pm

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Dec 29, 2012 2:00 pm

I was able to connect to the devices via winbox and some configuration on one of the costumers firewall (they both reside on their WAN-sides).
So here is more info, just gets me more confused. Don't find any info that leads me to the problem

The routerboards are named TC-SW-11 and TC-SW-12

Config on TC-SW-11
[admin@TC-SW-11] > export compact 
# jan/09/1970 23:53:22 by RouterOS 5.22
# software id = 92FB-FLPB
#
/interface bridge
add l2mtu=1594 name=vlan-eth1
add l2mtu=1594 name=vlan-eth2
add l2mtu=1594 name=vlan-eth3
add l2mtu=1594 name=vlan-eth4
/interface vlan
add interface=sfp1 l2mtu=1594 name=vlan2001-eth1 vlan-id=2001
add interface=sfp1 l2mtu=1594 name=vlan2002-eth2 vlan-id=2002
add interface=sfp1 l2mtu=1594 name=vlan2003-eth3 vlan-id=2003
add interface=sfp1 l2mtu=1594 name=vlan2004-eth4 vlan-id=2004
/ip hotspot user profile
set [ find default=yes ] idle-timeout=none keepalive-timeout=2m
/interface bridge port
add bridge=vlan-eth1 interface=vlan2001-eth1
add bridge=vlan-eth1 interface=ether1
add bridge=vlan-eth2 interface=vlan2002-eth2
add bridge=vlan-eth2 interface=ether2
add bridge=vlan-eth3 interface=vlan2003-eth3
add bridge=vlan-eth3 interface=ether3
add bridge=vlan-eth4 interface=ether4
add bridge=vlan-eth4 interface=vlan2004-eth4
/ip neighbor discovery
set vlan2001-eth1 disabled=yes
set vlan2002-eth2 disabled=yes
set vlan2003-eth3 disabled=yes
set vlan2004-eth4 disabled=yes
/lcd
set enabled=no
/system identity
set name=TC-SW-11
Config on TC-SW-12
[admin@TC-SW-12] > export compact 
# jan/02/1970 01:14:34 by RouterOS 5.22
# software id = TG8N-H60E
#
/interface bridge
add l2mtu=1594 name=vlan-eth1
add l2mtu=1594 name=vlan-eth2
add l2mtu=1594 name=vlan-eth3
add l2mtu=1594 name=vlan-eth4
/interface vlan
add interface=sfp1 l2mtu=1594 name=vlan2001-eth1 vlan-id=2001
add interface=sfp1 l2mtu=1594 name=vlan2002-eth2 vlan-id=2002
add interface=sfp1 l2mtu=1594 name=vlan2003-eth3 vlan-id=2003
add interface=sfp1 l2mtu=1594 name=vlan2004-eth4 vlan-id=2004
/ip hotspot user profile
set [ find default=yes ] idle-timeout=none keepalive-timeout=2m
/interface bridge port
add bridge=vlan-eth1 interface=ether1
add bridge=vlan-eth1 interface=vlan2001-eth1
add bridge=vlan-eth2 interface=vlan2002-eth2
add bridge=vlan-eth2 interface=ether2
add bridge=vlan-eth3 interface=ether3
add bridge=vlan-eth3 interface=vlan2003-eth3
add bridge=vlan-eth4 interface=ether4
add bridge=vlan-eth4 interface=vlan2004-eth4
/ip neighbor discovery
set vlan2001-eth1 disabled=yes
set vlan2002-eth2 disabled=yes
set vlan2003-eth3 disabled=yes
set vlan2004-eth4 disabled=yes
/lcd
set enabled=no
/system identity
set name=TC-SW-1
The log on the device that didn't get rebooted shows no info at all on the matter during the problem.
Not a single drop, collision or error on the RB that didn't get rebooted
[admin@TC-SW-11] > interface print stats-detail 
Flags: D - dynamic, X - disabled, R - running, S - slave 
 0  R  name="sfp1" rx-byte=307 662 466 591 tx-byte=337 942 548 458 rx-packet=248 027 252 tx-packet=266 464 321 rx-drop=0 
       tx-drop=0 rx-error=0 tx-error=0 

 1  R  name="ether1" rx-byte=3 190 205 512 tx-byte=27 782 537 317 rx-packet=14 340 881 tx-packet=24 907 349 rx-drop=0 
       tx-drop=0 rx-error=0 tx-error=0 

 2  R  name="ether2" rx-byte=333 140 775 437 tx-byte=279 405 224 063 rx-packet=252 008 656 tx-packet=223 121 811 rx-drop=0 
       tx-drop=0 rx-error=0 tx-error=0 
 
User avatar
kolorasta
Member Candidate
Member Candidate
Posts: 299
Joined: Sun Jun 25, 2006 11:55 pm
Location: Argentina

Re: RB2011UAS stopped working during heavy traffic on SFP

Tue Jan 01, 2013 6:32 pm

I think I have a similar problem. I have a vlan between rb600 and a rb2011. ROS 5.22 on both. At random days (15 days, 2 days, 10 days, 1 day) my rb2011 hangs but i noticed it hangs on ports that belongs to the vlan bridge. I plugged into another port and I could access to the rb2011. By now the only solution I found was a watchdog to some device unrecheable during the vlan bridge freeze.

I have to find out if it is a problem related to ROS or to the RB2011.

Sorry for my poor english and I hope you can understand what I wanted to say.

Enviado desde mi GT-N7000 usando Tapatalk 2
 
Dobby
Member
Member
Posts: 400
Joined: Wed Jan 11, 2012 12:07 am
Location: Hogwarts

Re: RB2011UAS stopped working during heavy traffic on SFP

Fri Jan 04, 2013 6:14 pm

Deleted because not related.
Last edited by Dobby on Mon Mar 11, 2013 2:15 am, edited 1 time in total.
 
User avatar
kolorasta
Member Candidate
Member Candidate
Posts: 299
Joined: Sun Jun 25, 2006 11:55 pm
Location: Argentina

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 10:38 am

With an rb493 will I be able to do the job?

Enviado desde mi GT-N7000 usando Tapatalk 2
 
Dobby
Member
Member
Posts: 400
Joined: Wed Jan 11, 2012 12:07 am
Location: Hogwarts

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 1:18 pm

Deleted because not related.
Last edited by Dobby on Mon Mar 11, 2013 2:14 am, edited 1 time in total.
 
CelticComms
Forum Guru
Forum Guru
Posts: 1766
Joined: Wed May 02, 2012 5:48 am

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 3:10 pm


@nissandata
.....................

If you have inside of one box setted up VLANs it should work proper, but if you are setting up an uplink and you will now
transport the VLANs through this uplinks it is so called out a "trunk" and this is purely not supported by the RB2011.
Or better and shorter by the on the board soldered Switch chips!

@kolorasta
The same as in the situation above, the RB2011 is not capable to build "trunks" or VLANs over more then one devices!
Why would it not be possible to operate VLAN trunks with the RB2011UAS? I have VLAN trunks running on both fibre and copper from RB2011UAS units without problems by simply creating the relevant VLAN interfaces and bridges within RouterOS.
Interlynx | Networking and Information Security Consultants & Trainers | Email: routerlynx@gmail.com
BGP | EIGRP | OSPF | MPLS | Firewall | VPN | IPsec | Multicast | QOS | IPv4/6 | STP | VLAN | PON | AE | M2M | and more!

 
Dobby
Member
Member
Posts: 400
Joined: Wed Jan 11, 2012 12:07 am
Location: Hogwarts

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 5:28 pm

Deleted because not related.
Last edited by Dobby on Mon Mar 11, 2013 2:12 am, edited 1 time in total.
 
CelticComms
Forum Guru
Forum Guru
Posts: 1766
Joined: Wed May 02, 2012 5:48 am

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 6:49 pm

That conversation probably related to differences between the chipsets on the routerboards. It is still possible to create VLAN "trunks" in RouterOS on the RB2011 but currently not possible to do so at the chip level on the AR8327.
Interlynx | Networking and Information Security Consultants & Trainers | Email: routerlynx@gmail.com
BGP | EIGRP | OSPF | MPLS | Firewall | VPN | IPsec | Multicast | QOS | IPv4/6 | STP | VLAN | PON | AE | M2M | and more!

 
Dobby
Member
Member
Posts: 400
Joined: Wed Jan 11, 2012 12:07 am
Location: Hogwarts

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 8:50 pm

Deleted because not related.
Last edited by Dobby on Mon Mar 11, 2013 2:12 am, edited 1 time in total.
 
User avatar
normis
MikroTik Support
MikroTik Support
Posts: 24268
Joined: Fri May 28, 2004 11:04 am
Location: Riga, Latvia

Re: RB2011UAS stopped working during heavy traffic on SFP

Sat Jan 05, 2013 10:12 pm

v6rc6 and v5.23 has this fixed, please try rc6 and see if your SFP problem still happens
No answer to your question? How to write posts

Who is online

Users browsing this forum: MSN [Bot] and 104 guests