miniRouter's ethernet ports die after random interval

I have a miniRouter that normally works fine, but every few hours its ethernet ports will go into a half-baked state and the box will need to be rebooted. The configuration is fairly simple:

        
    DSL modem -- (ether2)
                         \
                          -- (bonded interface) miniRouter (ether1) -- LAN
                         /
    DSL modem -- (ether3)

I have experimented with two software versions (2.9.43 and 2.9.48). In 2.9.43 the ethernet ports go into a 10Mbps/half duplex/up state (as does the LAN switch), yet the miniRouter will not respond to pings on ether1. In 2.9.48 the ports are all down.

I have tried two different miniRouters, three power supplies. The same thing happens with both of them. I have tried three different switches on ether1 (all Netgear, one 10/100 and 2 GigE), and various configurations of filtering and NAT. Even with the bare bones configuration (just routing, no filtering, no NAT, nothing vaguely optional) the ports will fail.

This post is my last ditch attempt to use the miniRouter for this application; any ideas?

Thanks!

Some details of the configuration would be useful. Bit hard to help when all we really have to go on is ‘It doesn’t work’.

Regards

Andrew

Well, the ethernet ports ought to always stay up, but that’s just my opinion. :smiley:

There’s a ton of stuff in the configuration; here’s what seems to be the relevant. This is from a miniRouter running 2.9.48. The failure mode on 2.9.48 is that the ethernet ports stop working and all the lights on all the ports go out. The switch attached to ether1 reports the link is down.

Let me know if anything else would be helpful.
thanks,
bri


/ interface bonding
add name=“bonding1” mtu=1500 arp=enabled slaves=ether2,ether3 mode=balance-rr primary=none link-monitoring=arp arp-interval=10s arp-ip-targets=172.17.0.1
mii-interval=100ms down-delay=0s up-delay=0s lacp-rate=30secs comment=“” disabled=no
/ interface ethernet
set ether1 name=“ether1” mtu=1500 mac-address=00:0C:42:13:35:6B arp=enabled disable-running-check=yes auto-negotiation=yes full-duplex=yes
cable-settings=default speed=100Mbps comment=“” disabled=no
set ether2 name=“ether2” mtu=1500 mac-address=00:0C:42:13:35:6C arp=enabled disable-running-check=yes auto-negotiation=yes full-duplex=yes
cable-settings=default speed=100Mbps comment=“” disabled=no
set ether3 name=“ether3” mtu=1500 mac-address=00:0C:42:13:35:6D arp=enabled disable-running-check=yes auto-negotiation=yes full-duplex=yes
cable-settings=default speed=100Mbps comment=“” disabled=no
/ ip address
add address=198.144.205.97/29 network=198.144.205.96 broadcast=198.144.205.103 interface=ether1 comment=“” disabled=no
add address=172.17.1.3/24 network=172.17.1.0 broadcast=172.17.1.255 interface=bonding1 comment=“” disabled=no
add address=172.17.0.3/24 network=172.17.0.0 broadcast=172.17.0.255 interface=bonding1 comment=“” disabled=no
add address=172.17.2.3/24 network=172.17.2.0 broadcast=172.17.2.255 interface=bonding1 comment=“” disabled=no
/ ip route
add dst-address=0.0.0.0/0 gateway=172.17.0.1 pref-src=172.17.0.3 distance=1 scope=255 target-scope=10 comment=“” disabled=no

From the bonding mini-howto:

  1. Which switches/systems does it work with?

In round-robin mode, it works with systems that support trunking:

  • Cisco 5500 series (look for EtherChannel support).
  • SunTrunking software.
  • Alteon AceDirector switches / WebOS (use Trunks).
  • BayStack Switches (trunks must be explicitly configured). Stackable
    models (450) can define trunks between ports on different physical
    units.
  • Linux bonding, of course !

In Active-backup mode, it should work with any Layer-II switches.

What configuration do you have on your switch?

Regards

Andrew

In this case, the DSL modems each have an ethernet connection directly to a port on the miniRouter. DSL-1 goes to ether2, and DSL-2 goes to ether3, and those two are bonded to bonding1, the WAN interface. There’s no switch involved there – I just plugged the modems in to the MT directly.

On ether1 (the LAN interface) I’ve tried three different ethernet switches: an old 10/100 Netgear (FS104), a small unmanaged Netgear GigE switch (GS105), and a managed Netgear switch (GSM7224).

You’re trying to load balance two DSL connections?

If so, this isn’t the way to do it; bonding is a layer 2 function. The wiki has some details on how failover & load balancing should be configured.

http://wiki.mikrotik.com/wiki/Routing

Regards

Andrew

try fixing to 100/half no auto to test

I had this problem

It looks to me like we have the bonding set up correctly, at least according to the wiki you cited. When it’s running it works very well, and we definitely get the throughput you’d expect from having the bonding work correctly. I just got a response from support, saying:

From what you wrote i understand that you are running RB150 in 10Mbit mode.
All 100 series boards have issues when running in 10Mbit mode. You can try
RB500, this board don’t have such issues.

Since the two DSL modems in my configuration are each running 10Mbs/half-duplex it could be that it’s causing the problem. However, my ISP has deployed a couple of similar configurations and I’m the only one having the problem.

I’m glad someone else is having this problem.

We have a RB150 miniRouter deployed at a client’s location as their main router. It has a single Internet connection on ether1, a Netgear WAP on ether2 with two VLANs and the rest of the ports (and one of the VLANs) are bridged together for the main office LAN. We have a hotspot configured on the second VLAN for public Internet access (so the user has to input a username and password).

Their network is a simple small business setup with about 8 Windows XP computers; the router had been deployed just fine for about 2 months. The problem started when we decided to put our Kaseya remote administration utility on their server.

This utility does create a lot of traffic when it is first installed as it audits everything on the computer and reports back to our central server (over the Internet). It also does some basic scans on the network.

The miniRouter’s Ethernet lights would ALL go out, but not the blue power light, and it would only be solved by a restart. Using a serial cable I was able to access the terminal interface so the router was not completely locked up (very strange). After installing Kaseya on their server the miniRouter would stop working constantly every 2-3 minutes.

I was monitoring system resources of the device but nothing peaked or showed any anomalies when the problem happened. After some troubleshooting I disabled the hotspot function and the miniRouter stayed up. I thought this was the problem but two days later we had the same problem again.

This seems like a bug in the hardware or how this platform using the RouterOS system. I know we are not over using the resources, and it seems that just the Ethernet ports shut off.

Hopefully Mikrotik sees this post and can help us look into this problem. If anyone else has had these issues please post.

Thanks for any help.

I had this exact thing, I changed the switch it was plugged into and it fine now.

Do you mean you changed the switch, the port on the router, or the port on the switch?

It seems like there must be something wrong with the routerBoard if this is happening at all.

I changed the HP procurve it was plugged into for a Cisco