Community discussions

MikroTik App
 
Andale
just joined
Topic Author
Posts: 5
Joined: Mon Jan 26, 2015 3:57 pm

6.41 bridge problem

Sat Jan 20, 2018 6:40 pm

Hi, I have an RB3011 on my lan, before 6.41 it was all working without problems... after upgrading from 6.40 to 6.41 I have this problem:

ETH1 is WAN (PPPoE Client) - still working
ETH2 is 1st lan connected to an HP switch and an AP: RB3011---[hp switch (not managed)]---[AP Mikrotik wAP 6.41 in Bridge mode] WORKING
ETH3 is 2nd lan (same subnet) connected to an HP switch and an AP: RB3011---[dlink switch (not managed)]---[AP Mikrotik mAP lite 6.41 in Bridge mode] NOT WORKING CORRECTLY

Clients connected to the mAP have now extremely slow internet connection. From mAP terminal there is a packet loss of 99% pinging 8.8.8.8 and the DNS is not working. If I connect the mAP to the HP switch, it works perfectly. From the mAP (on ETH3) I can ping without problems all internal IPs (also if connected to ETH2).

The map has 1bridge with ethernet (root) and wlan associated. If I connect the mAP on the ETH2 then DHCP releases the IP to the ethernet mac of the mAP. Connecting on the ETH3 the IP is associated to the wlan mac-address of the mAP.

The strangest thing is that if I open form winbox the ETH3 properties and I click "TORCH"... when torching the mAP starts pinging 8.8.8.8, DNS works, internet is ok.

ETH2 is the root-port of bridge1, ETH3,4,5 are designated ports. With 6.40 (switch config) all was working.

Now I have downgraded to 6.39.3 and I reloaded the backup and all is working... so we can exclude hardware problems or misconfiguration of the mAP.
 
User avatar
acruhl
Member
Member
Posts: 363
Joined: Fri Jul 03, 2015 7:22 pm

Re: 6.41 bridge problem

Sun Jan 21, 2018 12:12 am

As you know, 6.41 has a different bridge implementation so it's possible that whatever assumptions the code is making about creating the bridge and ports is wrong.

The fact that you're pinging and dropping packets seems like a layer 2 loop to me.

Although when you said you enabled torch and things started working, this might imply a routing loop, and not sure what 6.41 would have to do with that.

When you were losing pings to 8.8.8.8, was the error TTL timeout by chance?

You really should diagram it and make it clear if your APs are really APs or routers.
Stuff.
 
User avatar
acruhl
Member
Member
Posts: 363
Joined: Fri Jul 03, 2015 7:22 pm

Re: 6.41 bridge problem

Sun Jan 21, 2018 12:15 am

Wait, I said something wrong, if 8.8.8.8 works at all then it's probably not a routing loop problem, at least for the source subnet you are pinging from. Could be a routing loop for another one though.

It's not clear from the description, a diagram and some configs might be needed. Probably you should open a bug, but support is going to want to see something clean and easy to recreate.
Stuff.
 
User avatar
acruhl
Member
Member
Posts: 363
Joined: Fri Jul 03, 2015 7:22 pm

Re: 6.41 bridge problem

Sun Jan 21, 2018 12:27 am

And I reply to myself yet again.

When you have the ping issue, get a quick packet sniffer trace from the upstream interface and look at it in Wireshark. It should be obvious if it's a layer 2 or layer 3 loop at that point.

If it's layer 2, then the upgrade either made a wrong decision, or an incompatible decision.
Stuff.
 
Andale
just joined
Topic Author
Posts: 5
Joined: Mon Jan 26, 2015 3:57 pm

Re: 6.41 bridge problem

Sun Jan 21, 2018 6:07 pm

As you know, 6.41 has a different bridge implementation so it's possible that whatever assumptions the code is making about creating the bridge and ports is wrong.

The fact that you're pinging and dropping packets seems like a layer 2 loop to me.

Although when you said you enabled torch and things started working, this might imply a routing loop, and not sure what 6.41 would have to do with that.

When you were losing pings to 8.8.8.8, was the error TTL timeout by chance?

You really should diagram it and make it clear if your APs are really APs or routers.
I don't think that this is a loop problem. Ping 8.8.8.8 gives timeout error.

APs have same config and are simple AP bridge. The RB has ip 192.168.20.1 on the bridge1 interface. The AP has 192.168.20.11 and 192.168.20.12 on each bridge interface. ETH and WLAN are in the bridge.
The lan on eth2 and eth3 was are now in switch with 6.39.3 and all is working.

With 6.40 and bridge config (eth2-5 on the bridge1):

From 192.168.20.11 AP terminal (connected to ETH2) ping on 192.168.20.1 OK - ping on 8.8.8.8 OK - internet from clients OK
From 192.168.20.12 AP terminal (connected to ETH3) ping on 192.168.20.1 OK - ping on 8.8.8.8 KO - internet from clients KO - ping to any 192.168.20.x IP connected to eth2 OK
 
Andale
just joined
Topic Author
Posts: 5
Joined: Mon Jan 26, 2015 3:57 pm

Re: 6.41 bridge problem

Sun Jan 21, 2018 6:17 pm

I will try a new config from scratch. I'm afraid that the auto conversion from switch to bridge config intruduces some errors.
Sone firewall and nat rules needed to be corrected, so I'll try to reconfig from zero.

I'll post the results asap.
 
r0ute0ne
just joined
Posts: 2
Joined: Tue Sep 19, 2017 5:57 pm
Location: BIH / Sarajevo

Re: 6.41 bridge problem

Thu Feb 01, 2018 5:12 pm

I will try a new config from scratch. I'm afraid that the auto conversion from switch to bridge config intruduces some errors.
Sone firewall and nat rules needed to be corrected, so I'll try to reconfig from zero.

I'll post the results asap.
Did you manage to fix the problem?

Who is online

Users browsing this forum: MSN [Bot] and 84 guests