@terramoto said in LAN + VLAN on same interface:

@johnpoz

The tag, untag would have to be done at each powerline node.

Nope it would be done off a switch the other side of the powerline device.

Like this, just imaging the AP hanging off switch-3:-

image.png

Ah the powerline devices are doing the Wi-Fi as well, you need access-points.

@lonnie Yeah, that's one of the reasons we don't typically recommend bridges.

@fgervais
Yes firewall rules on the interfaces or on the bridge or NAT rules affect this traffic.

Also consider the values of the tuneables
net.link.bridge.pfil_member
net.link.bridge.pfil_bridge

Interfaces/Assignments/Bridges/

Edit

IMG_1056.jpeg

@swemattias while I would agree with you on the mini its a bit off.

Switches are switches are switch, vlans are vlans are vlans.

The vlan is either tagged or it isn't.. When you connect a end user device to a port on a switch is is almost always untagged.. Unless you have gone into the OS of that device and specifically set it to understand a tag..

Almost always a end user device will be untagged on a switch port, with the port in that vlan, and the pvid set on the port that you want the traffic to be in.

@johnpoz
Yep, brand new smart switch.

Thank you for your time!

@sledge As long as your switches are managed you would just trunk the uplinks and they'll carry the vlans with no problem.

@steveits Thanks for the help Steve! Bridge worked :)

I've finally managed to get this fixed, thanks to a kind soul found on the Internet. I basically got schooled(again!) on layer 2 traffic and having an extra pair of eyes go through the firewall config, I found out what the problem was. I was basically trying to shoehorn VLAN traffic through the switch and causing a loop(even with loop prevention turned off). However, this was not affecting my regular traffic which made me continue to troubleshoot and assume that my configuration was correct.

Considering my requirement has been that VMs talk to each and gets update over the internet and nothing outside of these VLANs, I added another interface to pfsense(trunk port) and in pfsense, changed the VLANs to be going through the new interface, rather than still pushing it through the physical LAN which I was trying to do. I now get DHCP AND the machines are able to reach out to the internet.

Once I added the trunk network interface as an additional NIC, it showed up as a 3rd interface on pfsense which showed as vmx2

ef00ec88-22ea-4b6e-a5cb-a5cd24c95b2e-image.png

I used the third NIC to pass my VLAN traffic
97e9f5c0-4b2a-4482-8320-999d1e4bbdaf-image.png

Earlier, I had configured VLAN to be going vmx1, by letting the traffic go out through the LAN/Trust interface and then trying to get it back through the same port (since I didn't have another NIC free on ESXi). Now, all my VMs are getting the correct IP address range

Well I would have preferred not to go in the « advanced user only » box on my first day but sometimes you gotta run before you can walk:

ifconfig bridge0 addr

@fgervais Thanks! Than my problem is probably related to the Hyper-V vSwitch I was using.

@jarhead Are you actually seeing a problem or just a few counters and only when you reboot? Lots of things happen when a system is rebooted and not just on the pfSense side. Assuming you're connected to a switch which also will run through some link up/link down procedures. If I was only seeing a few error counters on reboot and then no further incrementing or problems, I would personally move on to something else.

You'll likely need to send screen shots of the interface assignment page and detail exactly what you are doing to attempt to reassign the VLAN to the physical if.

@martywise have you tried to reconfigure the native lan for the port connected to your pfsense box. You could make the native lan of the switch trunk port different than the rest of the switch so it doesn't pass data other ports having same native vlan

@bigups43 Yes.

The big advantage of stackable switches that that you have redundancy as you can run a single LAGG over multiple switches.

@soheil-amiri

Max throughput can be reached on a bare bone system. A VM will always add overhead.

What has been done with pfSense : see one of the many Youtube videos.
I guess it can't up-scale forever, and thats why TNSR was created.

Now it works, solutions were as you wrote not to use vlan but lan on j4125 firewall.

Thanks @bingo600 and the rest of you for the help.

@rcoleman-netgate As I said, I had not assigned VLAN 128 to any interface, but thanks for the advice. Getting 504 would still be a pretty bugged out way in that case though.

The issue is gone for me now. I tried rebooting before posting here, and it didn't help. But now I rebooted again and it seems to have gone away. ¯\_(ツ)_/¯

I've just done a write up about the challenges of debugging these scenarios for reference: https://www.contradodigital.com/2022/07/25/how-to-troubleshoot-ping-icmp-not-working/

@mc-amz "switches" only shows up on Netgate hardware that has a switch built in... such as the 1100, 2100, 3100 and 7100 systems.

My 7100:
a2df6481-6062-4d6c-9b55-1659260d39ea-image.png

@johnpoz 🤣 🤣 🤣 🤣