Dear guys,
1 of our RB3011UiAS is disconnecting clients from pppoe and reconnect them after about 30 min or some hours.
Analyzing on various sections, I have seen menu/switch/host with empty records 
(other ones are all full of connected MACs). Does someone can help me?
Regards
Please post your /interface ethernet export.
/interface ethernet
set [ find default-name=ether1 ] mac-address=
4C:5E:0C:E3:9C:64 name=ISP1
set [ find default-name=ether2 ] mac-address=
4C:5E:0C:E3:9C:65 name=ISP2
set [ find default-name=ether3 ] mac-address=
4C:5E:0C:E3:9C:66 name=ISP3
set [ find default-name=ether4 ] mac-address=
4C:5E:0C:E3:9C:67 name=ISP4
set [ find default-name=ether5 ] mac-address=
4C:5E:0C:E3:9C:68 name=ISP5
set [ find default-name=ether6 ] mac-address=
4C:5E:0C:E3:9C:69 name=ISP6
set [ find default-name=ether7 ] mac-address=
4C:5E:0C:E3:9C:6A name=ISP7
set [ find default-name=ether8 ] mac-address=
4C:5E:0C:E3:9C:6B name=ISP8
set [ find default-name=ether9 ] mac-address=4C:5E:0C:E3:9C:6C name=ISP9
set [ find default-name=ether10 ] mac-address=4C:5E:0C:E3:9C:6D name=LAN
poe-out=off
You are not utilizing the switch chips. Hosts table represents switch chips internal mac address to port mapping.
http://wiki.mikrotik.com/wiki/Manual:Switch_Chip_Features#Port_Switching
Yes, I know. On this RB I donāt have any master/slave function because not interested.
But in other RBs there is no problem as the topic (disconnecting pppoe clients every from 30min to 5 hours) with also no master/slave function enabled.
The problem is not related to switching. How many pppoe clients are configured on this device? Are they disconnected all at once?
There are about 250 clients on it. There is not a single event. Usually about 2/3 of them together.
Enable ppp debug logging and generate supout.rif file after these disconnects and send it to support@mikrotik.com. Maybe we can tell something more about your problem from these logs.
/system logging add topics=ppp,debug,!packet
Update! Enabling ālink downsā column, we can see about 3 disconnections per hour after last machine reboot.
Could that confirm an hardware failure?? Donāt look at CPU % it was a rare peak, usually it is 7-25% max
Thanks
Hi guys,
problem effectively was just known āport flapping on ports 6-10ā (from ROS 6.36.rc20 renamed switch 2, finally) you can see topics here http://forum.mikrotik.com/t/port-flapping-ether6-link-down-up-on-rb3011uias-rm/95195/1 and here http://forum.mikrotik.com/t/port-flapping-ether6-link-down-up-on-rb3011uias-rm/95195/1
We solved only through a new official OS by Netinstall http://wiki.mikrotik.com/wiki/Manual:Netinstall and after updating (as e-mail received from support) to last RC version. Only now there are residual re-login of customers group, but it appear without a specific cluster. But since we use UBNT Rocket at moment (https://www.ubnt.com/airmax/rocketm/) we need to investigate if there is caused by them or what other.
P.S. host section was empty because that ports on RB3011 are not software bridged nor setted as master/slave (as you can see on previous screenshot there are 9xISP and 1xLAN with PCC). We are sure now of no failure for it because tested on an older RB1100AHx2
Thanks
Sorry, guys. I thinkā¦it was solved, because after last setup of previous post, RB3011 was about 8h without problems.
Butā¦now about 1 down every 4 hours, but there is. With identical configuration on RB1100 on the same place, no problems.
Attached screens.
