My desired configuration:
ether1 : VLAN trunk for PVID 1, 10 and 11 where 1 should be untagged (default) and 10 and 11 tagged
ether5 : VLAN with PVID 10 (untagged) but also carrying PVID 11 tagged
My setup as I understood is required since 6.41:
I have no VLAN interfaces (I think they are not required anymore when vlan-filtering is used?)
Now with this setup when I connect ether1 to my regular (PVID=1) network I cannot access RouterOS anymore. The only reason I see is that some VLAN configuration seems wrong but I don’t see the mistake in my config.
in the list suggests that you have deleted from your configuration a bridge which previously existed and configuration of some member port still refers to it.
Other than that, is your IP address from vlan 1 (which you want to have tagless at ****
ether1
) attached to
bridge-vlan
as its carrying interface or to some other one?
I have no VLAN interfaces (I think they are not required anymore when vlan-filtering is used?)
You need interfaces of type ****
vlan
only if you want to run IP addresses in these VLANs. It has nothing to do with vlan-filtering. vlan-filtering only means that if you have multiple vlans on a single bridge as you do, you prevent frames with VLAN IDs unsupported on a given port from entering the bridge. Without vlan-filtering active, frames with any VLAN ID can enter the bridge.
Somehow there is always some weirdness when ****
pvid=1
is used, so maybe your trouble comes from there. If you don’t need VLAN ID 1 tagged anywhere, better use some other ID like 1111 for ether1’s pvid and create a
Thanks for your feedback.
I think I’m getting closer but still my basic testcase is failing.
I cannot connect to the device when connected via the ether1 interface from regular (non-vlan) network. This is the very basic thing which obviously needs to work.
Other than that my configuration now looks like:
For simplicity I left out the handling of vlan11 now until vlan1 and vlan10 works correctly.
The device IP is 192.168.250.251 on bridge-vlan and 192.168.250.250 on bridge/ether2 for now until the vlan bridge works.
/ip address
add address=192.168.250.250/24 comment=defconf interface=bridge network=192.168.250.0
add address=192.168.250.251/24 interface=bridge-vlan network=192.168.250.0
EDIT: Actually do I need to change anything so I can connect via winbox, webfig or similar to the device? Just seen interface list entries for mac-winbox → bridge etc. But bridge and bridge-vlan are currently disconnected.
So I think I’m getting there. vlan10 and vlan11 testing still missing but vlan1 works finally.
One thing I noticed though:
Using winbox I can connect to the IP of the vlan bridge but the device does not show up under neighbors which seems a bit strange to me. Is that to be expected?
I’m useless for Winbox related questions, I use it very rarely. But your two addresses from the same subnet attached to two interfaces in different VLANs are a voucher for trouble alone
Interesting. Actually I had indeed two IPs on the different bridges. But both bridges are supposed to work with VLAN1. So I didn’t expect a problem with that.
But then again removing one seems to have fixed my CAPSMAN connection issue but still not the issue that winbox does not detect the device still. I’m fine with using the IP but this smells like something is wrong still. I expect that bridge-vlan is still in the same network segment (for untagged traffic) so winbox should see it.
are totally isolated from each other at L2 unless you connect ether1 with ether2 using a cable. So the network stack at L3 has two gateways to that network, with the same priority (distance), so it can throw a coin to choose which one of them to use to send a packet. That’s not so much of an issue is the cable between ether1 and ether2 is there, but becomes a surprise generator when it is not.
