Hi all,
I'm troubleshooting a VLAN issue that appeared after upgrading to OPNsense 26.7.4 / FreeBSD 15.1.
My setup is roughly:
Cisco 2960X
|
Po3 LACP / 802.1Q
|
Intel X550 ix0 + ix1
|
lagg0
|
VLAN interfaces
|
OPNsense bridges
|
Kea DHCP
I also have an Intel i226 2.5 GbE trunk, with corresponding VLAN interfaces bridged to the X550/LAGG VLAN interfaces.
This affects all VLANs traversing the Cisco 2960X / Intel X550 LAGG path. VLAN 15 is simply the VLAN I am using for troubleshooting.
Clients on the Cisco side fail DHCP and remain on 169.254.x.x.
Cisco switch looks healthy: Po3 is UP, both LACP members are bundled, VLAN 15 is forwarding, and the test client MAC is correctly learned on its VLAN 15 access port.
The interesting part is simultaneous packet captures of the same DHCP transaction:
bridge1: DHCP REQUEST ✓ ACK ✓
vlan0.1.15: DHCP REQUEST ✓ ACK ✓
lagg0: DHCP not visible
ix0 / ix1: DHCP not visible
Kea therefore appears to receive the request and generate an ACK, but the client never receives/configures the lease.
I found FreeBSD bug 276936, describing packet-forwarding problems involving ixgbe, VLAN interfaces and bridges. The topology isn't identical, but it looks potentially relevant.
Has anyone experienced similar issues with OPNsense 26.7, Intel X550, LAGG/LACP, VLANs and bridges, or have suggestions for further diagnostics?
Any help or feedback would be greatly appreciated, as I am stuck until I can fix this, as my network relies on the VLAN Bridges to connect my home network
Thanks in advance
Hi Team
I have found that there is a bug, meaning I have found the root cause of the issue
The issue is that when you use Bridged VLANs, you cannot use Kea DHCP, it simply doesnt allow you to connect
When you use Bridged VLANS, you must use ISC Legacy DHCP, I have confirmed my connection issue is resolved
If anyone in the OPNsense forum knows how to report a bug, as I believe this is a bug in the system
Regards
Can you post the full output of "ifconfig", please? You can remove public IP addresses.
Hi Patrick
Thanks for replying to my post!
Please see the ifconfig output below. I've redacted a few personal details from the VLAN/interface descriptions, but the underlying configuration and topology remain unchanged, so hopefully it provides everything needed to understand the setup.
My ultimate goal is to move away from Legacy ISC DHCP and transition fully to Kea DHCP. That was actually one of the objectives when I rebuilt the router from scratch on OPNsense 26.7.4.
At the moment, the bridged VLAN configuration is working correctly with Legacy ISC DHCP, whereas I have been unable to achieve the same result with Kea.
Thanks again for taking the time to look at this.
----------------------------------------------------
root@OPNsense:~ # ifconfig
ix0: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
options=4803828<VLAN_MTU,JUMBO_MTU,WOL_UCAST,WOL_MCAST,WOL_MAGIC,HWSTATS,MEXTPG>
ether (redacted MAC)
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
ix1: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
options=4803828<VLAN_MTU,JUMBO_MTU,WOL_UCAST,WOL_MCAST,WOL_MAGIC,HWSTATS,MEXTPG>
ether (redacted MAC)
hwaddr (redacted MAC)
media: Ethernet autoselect (1000baseT <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
igc0: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
description: TRUNK (lan)
options=4902028<VLAN_MTU,JUMBO_MTU,WOL_MAGIC,NETMAP,HWSTATS,MEXTPG>
ether (redacted MAC)
inet 10.100.14.1 netmask 0xfffffff8 broadcast 10.100.14.7
groups: Interface_GROUP
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
igc1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu 1500
description: WAN (wan)
options=4802028<VLAN_MTU,JUMBO_MTU,WOL_MAGIC,HWSTATS,MEXTPG>
ether (redacted MAC)
inet6 (redacted IPv6)
groups: Interface_GROUP
media: Ethernet autoselect
status: no carrier
nd6 options=23<PERFORMNUD,ACCEPT_RTADV,AUTO_LINKLOCAL>
lo0: flags=1008049<UP,LOOPBACK,RUNNING,MULTICAST,LOWER_UP> metric 0 mtu 16384
options=680003<RXCSUM,TXCSUM,LINKSTATE,RXCSUM_IPV6,TXCSUM_IPV6>
inet 127.0.0.1 netmask 0xff000000
inet 10.100.100.1 netmask 0xffffffff
inet6 ::1 prefixlen 128
inet6 fe80::1%lo0 prefixlen 64 scopeid 0x5
groups: lo
nd6 options=21<PERFORMNUD,AUTO_LINKLOCAL>
enc0: flags=0 metric 0 mtu 1536
options=0
groups: enc
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
pflog0: flags=0 metric 0 mtu 33152
options=0
groups: pflog
pfsync0: flags=0 metric 0 mtu 1500
options=0
maxupd: 128 defer: off version: 1500
syncok: 1
groups: pfsync
lagg0: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
description: LAGG_INTERFACE (opt1)
options=4903828<VLAN_MTU,JUMBO_MTU,WOL_UCAST,WOL_MCAST,WOL_MAGIC,NETMAP,HWSTATS,MEXTPG>
ether (redacted MAC)
hwaddr 00:00:00:00:00:00
inet 10.100.23.1 netmask 0xfffffff8 broadcast 10.100.23.7
laggproto lacp lagghash l2,l3,l4
laggport: ix0 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
laggport: ix1 flags=1c<ACTIVE,COLLECTING,DISTRIBUTING>
groups: lagg
media: Ethernet autoselect
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
vlan0.1.10: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
description: (redacted VLAN 10 interface)
options=4000000<MEXTPG>
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 10 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
media: Ethernet autoselect
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
vlan0.2.10: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
description: (redacted VLAN 10 interface)
options=4000000<MEXTPG>
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 10 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
vlan0.1.15: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
description: (redacted VLAN 15 interface)
options=4000000<MEXTPG>
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 15 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
media: Ethernet autoselect
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
vlan0.2.15: flags=1008943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST,LOWER_UP> metric 0 mtu 1500
description: (redacted VLAN 15 interface)
options=4000000<MEXTPG>
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 15 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
nd6 options=29<PERFORMNUD,IFDISABLED,AUTO_LINKLOCAL>
vlan0.1.20:
description: (redacted VLAN 20 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 20 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
status: active
vlan0.2.20:
description: (redacted VLAN 20 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 20 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
vlan0.1.30:
description: (redacted VLAN 30 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 30 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
status: active
vlan0.2.30:
description: (redacted VLAN 30 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 30 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
vlan0.1.35:
description: (redacted VLAN 35 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 35 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
status: active
vlan0.2.35:
description: (redacted VLAN 35 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 35 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
vlan0.1.40:
description: (redacted VLAN 40 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 40 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
status: active
vlan0.2.40:
description: (redacted VLAN 40 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 40 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
vlan0.1.45:
description: (redacted VLAN 45 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 45 vlanproto: 802.1q vlanpcp: 0 parent interface: lagg0
status: active
vlan0.2.45:
description: (redacted VLAN 45 interface)
ether (redacted MAC)
groups: vlan (redacted VLAN groups)
vlan: 45 vlanproto: 802.1q vlanpcp: 0 parent interface: igc0
media: Ethernet autoselect (2500Base-T <full-duplex>)
status: active
bridge0:
description: BRIDGE_VLAN10_INTERFACE
ether (redacted MAC)
inet 10.100.10.1 netmask 0xffffff00 broadcast 10.100.10.255
proto rstp maxaddr 2000 timeout 1200
member: vlan0.2.10 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
port 11 priority 128 path cost 20000 vlan protocol 802.1q
member: vlan0.1.10 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
port 10 priority 128 path cost 55 vlan protocol 802.1q
groups: bridge (redacted bridge groups)
bridge1:
description: BRIDGE_VLAN15_INTERFACE
ether (redacted MAC)
inet 10.100.15.1 netmask 0xffffff00 broadcast 10.100.15.255
proto rstp maxaddr 2000 timeout 1200
member: vlan0.2.15 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
port 13 priority 128 path cost 20000 vlan protocol 802.1q
member: vlan0.1.15 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
port 12 priority 128 path cost 55 vlan protocol 802.1q
groups: bridge (redacted bridge groups)
bridge2:
description: BRIDGE_VLAN20_INTERFACE
ether (redacted MAC)
inet 192.168.50.1 netmask 0xffffff00 broadcast 192.168.50.255
proto rstp maxaddr 2000 timeout 1200
member: vlan0.2.20 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
member: vlan0.1.20 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
groups: bridge (redacted bridge groups)
bridge3:
description: BRIDGE_VLAN30_INTERFACE
ether (redacted MAC)
inet 10.100.30.1 netmask 0xffffff00 broadcast 10.100.30.255
member: vlan0.2.30 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
member: vlan0.1.30 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
groups: bridge (redacted bridge groups)
bridge4:
description: BRIDGE_VLAN35_INTERFACE
ether (redacted MAC)
inet 10.100.35.1 netmask 0xffffff00 broadcast 10.100.35.255
member: vlan0.2.35 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
member: vlan0.1.35 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
groups: bridge (redacted personal bridge groups)
bridge5:
description: BRIDGE_VLAN40_INTERFACE
ether (redacted MAC)
inet 10.100.40.1 netmask 0xffffff00 broadcast 10.100.40.255
member: vlan0.2.40 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
member: vlan0.1.40 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
groups: bridge (redacted personal bridge groups)
bridge6:
description: BRIDGE_VLAN45_INTERFACE
ether (redacted MAC)
inet 10.100.45.1 netmask 0xffffff00 broadcast 10.100.45.255
member: vlan0.2.45 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
member: vlan0.1.45 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>
groups: bridge (redacted personal bridge groups)
zen0:
flags=8010<POINTOPOINT,MULTICAST> metric 0 mtu 1500
groups: tun (redacted VPN group)
---------------------------------------------------------------------
Alright. You VLAN and bridge configuration is correct. You assigned the OPNsense interfaces to the bridges and there are no IP addresses on any bridge members.
I assume you enabled Kea on the bridge interfaces, too? Can't do anything but use the assigned symbolic names in the UI if I am not mistaken. Are you trying with raw or UDP sockets? Should be raw, IMHO.
Kind regards,
Patrick
Hi Patrick,
Thanks for confirming that the VLAN/bridge configuration itself looks correct.
Yes, when I was originally testing Kea, I configured the DHCPv4 subnets against the assigned OPNsense bridge interfaces, not against the individual VLAN bridge members.
For example, VLAN 15 is effectively:
lagg0 -> vlan0.1.15 ---\
bridge1 -> 10.100.15.1/24
igc0 -> vlan0.2.15 ---/
The IP address/gateway resides only on bridge1. Neither vlan0.1.15 nor vlan0.2.15 has an IP address.
With Kea enabled, the client would send DHCP requests but would ultimately fail to obtain/configure an address.
To eliminate Kea from the equation, I then performed the following test:
Disabled the relevant Kea DHCP interfaces/subnets.
Stopped/disabled Kea so that it was no longer servicing DHCP.
Enabled Legacy ISC DHCPv4 on the assigned bridge interface.
Configured the same subnet, gateway and DHCP pool on the bridge.
Renewed DHCP from the same client, through the same Cisco switch, LAGG, VLAN and OPNsense bridge topology.
With ISC DHCP, the client successfully received a lease:
Client: 10.100.15.10/24
Gateway: 10.100.15.1
DNS: 10.100.100.1
I also tested the client with a static address (10.100.15.150/24) beforehand and confirmed bidirectional connectivity to the bridge gateway at 10.100.15.1.
So at this point the same underlying bridge/VLAN topology works for normal IP traffic and successfully provides DHCP when using Legacy ISC DHCP. The failure appears when I use Kea instead.
Regarding raw versus UDP sockets, I did not intentionally change Kea away from its default socket behaviour. My understanding is that Kea DHCPv4 uses raw sockets by default, unless dhcp-socket-type is explicitly configured otherwise. ISC's Kea documentation also notes that UDP sockets have limitations when responding directly to clients that do not yet have an IPv4 address.
I have a backup configuration on my test Virtual firewall, I can produce this if required.
Kind regards,
Stefan
Can you show the Kea configuration? Possibly I can spot something "wrong"?
Also, but this is probably not related to the problem - with your advanced use of bridges I'd recommend creating and setting this tunable:
tunable: net.link.bridge.inherit_mac
value: 1
This way the bridges have a static MAC address across reboots.
Hi Patrick,
Thanks, that would be greatly appreciated.
I can provide the generated Kea DHCPv4 configuration. Before I switched over to Legacy ISC for testing, Kea was configured against the assigned bridge interfaces themselves, not the individual VLAN members.
For example, VLAN 15 is:
lagg0 -> vlan0.1.15 ---\
bridge1 -> 10.100.15.1/24
igc0 -> vlan0.2.15 ---/
Neither VLAN member has an IP address. The IPv4 address exists only on bridge1.
For the A/B test I completely disabled the Kea DHCP interfaces/service before enabling Legacy ISC DHCPv4 on the bridge. With essentially the same subnet, pool, gateway and DNS configuration, ISC successfully leases addresses across the bridge. The same Mac that previously failed with Kea immediately obtained 10.100.15.10/24 via ISC and could communicate normally.
So at present the evidence is:
Same client
Same switch port
Same Cisco trunk/LACP
Same VLAN interfaces
Same bridge
Same OPNsense gateway
Kea DHCP -> fails
ISC DHCP -> works
Regarding the socket type, I did not deliberately change Kea to UDP sockets. Kea normally uses raw sockets by default, unless dhcp-socket-type is explicitly configured otherwise.
I'll retrieve the actual generated Kea configuration so we can confirm exactly:
"interfaces-config": {
"interfaces": [...],
"dhcp-socket-type": "..."
}
Thanks for that suggestion. I hadn't configured it.
I checked the FreeBSD 15.1 bridge documentation. A non-zero net.link.bridge.inherit_mac causes a newly created bridge to inherit the MAC address of its first member, providing a more predictable bridge MAC. However, FreeBSD currently describes this feature as experimental, disabled by default, and notes that it can break some L2 protocols such as PPPoE.
Given that, I'll certainly investigate it, but I'd prefer to keep it separate from the DHCP troubleshooting initially so that I don't introduce another variable while comparing Kea against ISC.
Kind regards,
Stefan
Quote from: merrins63 on September 29, 2026, 11:27:46 AMI can provide the generated Kea DHCPv4 configuration.
So please do :-)
Hi Patrick,
Absolutely. I managed to retrieve the actual generated Kea DHCPv4 configuration from the firewall. I've only redacted the internal domain name, UUIDs and descriptive labels. The technically relevant configuration is unchanged.
This also confirms your earlier question regarding socket type. Kea was configured with:
"dhcp-socket-type": "raw"
Please see the below extract from my Test Firewall
________________________________________________
{
"Dhcp4": {
"valid-lifetime": 4000,
"decline-probation-period": 600,
"interfaces-config": {
"interfaces": [
"bridge0",
"bridge1",
"bridge2",
"bridge3",
"bridge4",
"bridge5",
"bridge6",
"vtnet1",
"lagg0",
"vtnet4"
],
"dhcp-socket-type": "raw",
"service-sockets-max-retries": 5,
"service-sockets-retry-wait-time": 5000
},
"lease-database": {
"type": "memfile",
"persist": true
},
"control-socket": {
"socket-type": "unix",
"socket-name": "/var/run/kea/kea4-ctrl-socket"
},
"loggers": [
{
"name": "kea-dhcp4",
"output_options": [
{
"output": "syslog"
}
],
"severity": "INFO"
}
],
"subnet4": [
{
"id": 1,
"subnet": "10.100.100.0/24",
"next-server": "",
"match-client-id": true,
"option-data": [
{
"name": "domain-name-servers",
"data": "10.100.100.1"
},
{
"name": "routers",
"data": "10.100.100.1"
},
{
"name": "domain-name",
"data": "(redacted internal domain)"
}
],
"pools": [
{
"pool": "10.100.100.10-10.100.100.20"
}
],
"reservations": [],
"valid-lifetime": 86400,
"user-context": {
"uuid": "(redacted UUID)",
"description": "(redacted subnet description)"
}
},
{
"id": 2,
"subnet": "10.100.14.1/29",
"next-server": "",
"match-client-id": true,
"option-data": [
{
"name": "domain-name-servers",
"data": "10.100.100.1"
},
{
"name": "routers",
"data": "10.100.14.1"
},
{
"name": "domain-name",
"data": "(redacted internal domain)"
}
],
"pools": [
{
"pool": "10.100.14.2-10.100.14.6"
}
],
"reservations": [],
"valid-lifetime": 86400,
"user-context": {
"uuid": "(redacted UUID)",
"description": "(redacted subnet description)"
}
},
{
"id": 3,
"subnet": "10.100.10.0/24",
"next-server": "",
"match-client-id": true,
"option-data": [
{
"name": "domain-name-servers",
"data": "10.100.100.1"
},
{
"name": "routers",
"data": "10.100.10.1"
},
{
"name": "domain-name",
"data": "(redacted internal domain)"
}
],
"pools": [
{
"pool": "10.100.10.10-10.100.10.250"
}
],
"reservations": [],
"valid-lifetime": 86400,
"user-context": {
"uuid": "(redacted UUID)",
"description": "(redacted VLAN 10 description)"
}
},
{
"id": 4,
"subnet": "10.100.15.0/24",
"next-server": "",
"match-client-id": true,
"option-data": [
{
"name": "domain-name-servers",
"data": "10.100.100.1"
},
{
"name": "routers",
"data": "10.100.15.1"
},
{
"name": "domain-name",
"data": "(redacted internal domain)"
}
],
"pools": [
{
"pool": "10.100.15.10-10.100.15.250"
}
],
"reservations": [],
"valid-lifetime": 86400,
"user-context": {
"uuid": "(redacted UUID)",
"description": "(redacted VLAN 15 description)"
}
},
{
"id": 5,
"subnet": "10.100.23.1/29",
"next-server": "",
"match-client-id": true,
"option-data": [
{
"name": "domain-name-servers",
"data": "10.100.100.1"
},
{
"name": "routers",
"data": "10.100.23.1"
},
{
"name": "domain-name",
"data": "(redacted internal domain)"
}
],
"pools": [
{
"pool": "10.100.23.2-10.100.23.6"
}
],
"reservations": [],
"valid-lifetime": 86400,
"user-context": {
"uuid": "(redacted UUID)",
"description": "(redacted subnet description)"
}
}
],
"hooks-libraries": [
{
"library": "/usr/local/lib/kea/hooks/libdhcp_lease_cmds.so"
},
{
"library": "/usr/local/lib/kea/hooks/libdhcp_host_cmds.so"
}
]
}
}
_______________________________________________________________________
Regards
[...]
"subnet": "10.100.14.1/29",
[...]
"subnet": "10.100.23.1/29",
[...]
These are not valid subnets. 10.100.14.0/29 and 10.100.23.ü/29 would be if that is what is intended.
Thanks, good spot.
Those two /29 entries are unrelated to the issue I'm troubleshooting. They are temporary device/transit networks on this test firewall and were not involved in the DHCP testing.
The relevant test networks are:
VLAN 10
Subnet: 10.100.10.0/24
Gateway: 10.100.10.1
Bridge: bridge0
VLAN 15
Subnet: 10.100.15.0/24
Gateway: 10.100.15.1
Bridge: bridge1
VLAN 15 has been my primary troubleshooting network. With the same bridge configuration and client:
Kea DHCP -> client fails to obtain/use a lease
ISC DHCP -> client successfully receives 10.100.15.10/24
Static IP -> connectivity to 10.100.15.1 works
So please focus primarily on the 10.100.10.0/24 and 10.100.15.0/24 Kea configuration.
I'll correct the two /29 definitions separately. Thanks for spotting them
These definitions match the interface configuration. I am out of ideas ... er ... one moment ...
Did you check this option?
(https://forum.opnsense.org/index.php?action=dlattach;attach=58356;image)
If you did not that would perfectly explain your observed symptoms.
ISC creates the necessary firewall rules for DHCP to work by default, if I remember correctly. For Kea that's optional, because some users complained that OPNsense should not create any rules automatically giving full control to the admin.
I have got a question that really bugs me :
Why would you do this =>
Quote from: merrins63 on September 28, 2026, 01:38:52 PMI also have an Intel i226 2.5 GbE trunk, with corresponding VLAN interfaces bridged to the X550/LAGG VLAN interfaces.
Aren't you effectively Bridging 2x 2,5 Gbps with 2 x 10 Gbps ?!
What is the purpose of such a setup ??
Quote from: Patrick M. Hausen on September 29, 2026, 01:20:12 PMThese definitions match the interface configuration. I am out of ideas ... er ... one moment ...
Did you check this option?
(https://forum.opnsense.org/index.php?action=dlattach;attach=58356;image)
If you did not that would perfectly explain your observed symptoms.
ISC creates the necessary firewall rules for DHCP to work by default, if I remember correctly. For Kea that's optional, because some users complained that OPNsense should not create any rules automatically giving full control to the admin.
Quote from: Patrick M. Hausen on September 29, 2026, 01:20:12 PMThese definitions match the interface configuration. I am out of ideas ... er ... one moment ...
Did you check this option?
(https://forum.opnsense.org/index.php?action=dlattach;attach=58356;image)
If you did not that would perfectly explain your observed symptoms.
ISC creates the necessary firewall rules for DHCP to work by default, if I remember correctly. For Kea that's optional, because some users complained that OPNsense should not create any rules automatically giving full control to the admin.
Yes this setting was already enabled by default. I turned it off, applied the settings then re enabled it
It made no difference and behaviour using Kea DHCP still persists
Cheers
Quote from: nero355 on September 29, 2026, 03:17:58 PMI have got a question that really bugs me :
Why would you do this =>
Quote from: merrins63 on September 28, 2026, 01:38:52 PMI also have an Intel i226 2.5 GbE trunk, with corresponding VLAN interfaces bridged to the X550/LAGG VLAN interfaces.
Aren't you effectively Bridging 2x 2,5 Gbps with 2 x 10 Gbps ?!
What is the purpose of such a setup ??
The issue I'm trying to solve with this design is mainly about making the best use of the different network speeds available in my home.
I have a traditional 1 GbE wired network, which is perfectly adequate for many of my wired IoT devices. However, my wireless access points and newer switches are capable of 2.5 GbE.
I need my wired and wireless IoT devices to remain on the same VLANs and subnets, while allowing the 2.5 GbE side of the network to make use of the faster interfaces rather than forcing everything through the existing 1 GbE switching infrastructure.
I'm also planning to upgrade my Internet connection to 2000/200 in the next few months, and I already have several multi-gigabit devices that will be able to take advantage of that bandwidth.
Previously, I used the more traditional approach of a single 24-port Gigabit Ethernet switch. However, with affordable 2.5 GbE switches now readily available, I wanted to take advantage of the Cat6A cabling already installed throughout my house.
The important point is that this design isn't intended for redundancy. The objective is to utilise the available 1 GbE and 2.5 GbE links efficiently while keeping devices within their existing VLANs.
Using bridged VLAN interfaces on OPNsense allows me to achieve that design.
Hopefully that explains the reasoning behind my setup a little better.
Even if my post doesn't help you solve the specific problem, I wanted to write because I'm trying to understand your approach. You never stop learning, and I might get some ideas to rethink my own network design.
As I understand it, you are trying to get the most out of your existing 1G and 2.5G components. However, the current issue following the upgrade is a direct result of software bridging on the OPNsense.
In my opinion, using OPNsense purely as a bridge or Layer 2 switch fails to leverage its capabilities and doesn't align with its intended design.
Using the firewall as an L2 switch to connect these two worlds within the same VLANs creates a highly complex stack involving LACP, VLAN tagging, BSD bridges, and firewall rules.
Even if the problem is resolved, the solution remains fragile, places unnecessary load on the firewall CPU, and introduces extra latency.
To meet your bandwidth requirements, how about considering these solutions instead?
1. A central 2.5G/10G switch handles all L2 switching in hardware. This offloads the OPNsense, allowing it to connect to the network via a clean trunk.
or
2. Separating the 2.5G devices into their own subnets at Layer 3. This allows you to use OPNsense exactly as it was designed to be used.
Communication (mDNS, SSDP) between the networks can be ensured using a multicast relay, for example (https://forum.opnsense.org/index.php?topic=52705.0).
The FreeBSD bridge has been rewritten from scratch and works perfectly fine. There is nothing wrong with this design if implemented correctly.
To reiterate: My focus is on understanding the situation. There have already been several discussions on the forum regarding the use of OPNsense as a bridge.
And as I have learned, there are circumstances that make bridging necessary—though it remains only the second-best option.
Even though FreeBSD's bridge mode can work well, I do not understand why one would want to use it.
If you are building a network and planning network segmentation for valid reasons, the setup looks like this:
ISP – OPNsense Router – VLAN Trunk – Switch Segmentation – Clients in different network segments.
You could certainly make it more complex by separating the routing and firewall functions, but what advantage would that offer in the context of the described setup and requirements?
Sometimes it's just how your apartment or office is built, the location of the "Internet socket", and so on. E.g. years ago in the office of my wife there was a single network connection to a small cabinet on which the laser printer was placed and it was decided that the new FreeNAS was to be put next to the printer. So I bridged the two ports of the NAS to connect NAS and printer to the network via that single line - why not? Instead of messing with an extra switch, needing yet another power outlet etc.
The OP might tell us their reasons :-)
Quote from: merrins63 on September 29, 2026, 11:06:34 PMThe issue I'm trying to solve with this design is mainly about making the best use of the different network speeds available in my home.
I have a traditional 1 GbE wired network, which is perfectly adequate for many of my wired IoT devices.
However, my wireless access points and newer switches are capable of 2.5 GbE.
I need my wired and wireless IoT devices to remain on the same VLANs and subnets, while allowing the 2.5 GbE side of the network to make use of the faster interfaces rather than forcing everything through the existing 1 GbE switching infrastructure.
I'm also planning to upgrade my Internet connection to 2000/200 in the next few months, and I already have several multi-gigabit devices that will be able to take advantage of that bandwidth.
Previously, I used the more traditional approach of a single 24-port Gigabit Ethernet switch.
However, with affordable 2.5 GbE switches now readily available, I wanted to take advantage of the Cat6A cabling already installed throughout my house.
The important point is that this design isn't intended for redundancy.
The objective is to utilise the available 1 GbE and 2.5 GbE links efficiently while keeping devices within their existing VLANs.
Using bridged VLAN interfaces on OPNsense allows me to achieve that design.
Hopefully that explains the reasoning behind my setup a little better.
You are now talking about 1 Gbps and 2,5 Gbps connections, but the X550 is AFAIK a 10 Gbps NIC ?!
I am really curious what the complete setup will look like once you have upgraded everything as you are planning to, so if you feel like it let us know once everything is up and running :)
Quote from: nero355 on September 30, 2026, 05:29:56 PMQuote from: merrins63 on September 29, 2026, 11:06:34 PMThe issue I'm trying to solve with this design is mainly about making the best use of the different network speeds available in my home.
I have a traditional 1 GbE wired network, which is perfectly adequate for many of my wired IoT devices.
However, my wireless access points and newer switches are capable of 2.5 GbE.
I need my wired and wireless IoT devices to remain on the same VLANs and subnets, while allowing the 2.5 GbE side of the network to make use of the faster interfaces rather than forcing everything through the existing 1 GbE switching infrastructure.
I'm also planning to upgrade my Internet connection to 2000/200 in the next few months, and I already have several multi-gigabit devices that will be able to take advantage of that bandwidth.
Previously, I used the more traditional approach of a single 24-port Gigabit Ethernet switch.
However, with affordable 2.5 GbE switches now readily available, I wanted to take advantage of the Cat6A cabling already installed throughout my house.
The important point is that this design isn't intended for redundancy.
The objective is to utilise the available 1 GbE and 2.5 GbE links efficiently while keeping devices within their existing VLANs.
Using bridged VLAN interfaces on OPNsense allows me to achieve that design.
Hopefully that explains the reasoning behind my setup a little better.
You are now talking about 1 Gbps and 2,5 Gbps connections, but the X550 is AFAIK a 10 Gbps NIC ?!
I am really curious what the complete setup will look like once you have upgraded everything as you are planning to, so if you feel like it let us know once everything is up and running :)
Thanks for all the thoughtful feedback and curiosity about the design!
I completely agree that it's an unconventional layout, but it serves a very specific purpose given my current hardware constraints and future goals.
To clear up a few points regarding the setup and why I chose this path:
Hardware Constraints & Uplinks: My current core switch is a Cisco Catalyst 2960-X, which is limited to dual 1GbE links. Rather than ripping and replacing or reconfiguring my existing LAGG trunk setup on the X550 NICs, I utilized the 10GbE capability of the X550 to bridge directly against that existing infrastructure.
Future-Proofing: I am planning ahead for a future upgrade to a 2.5GbE managed switch that features 10GbE uplinks. This design lets me bridge the gap right now without having to completely redesign my VLANs and LAGG configuration later down the track.
Handling Software Bridge Overhead: I'm well aware that software bridging places extra demands on the system, which is precisely why I deliberately chose quite powerful hardware to run OPNsense. The firewall CPU handles the software bridge effortlessly.
Why It Works: As a network enthusiast, I ran into a limitation with my current gear, and software bridging elegantly overcame it while keeping my IoT devices cleanly structured within their existing VLANs across mixed-speed links. It's practical, it's cost-effective for my current phase, and up until this recent upgrade issue, it has performed fantastically.
Given that FreeBSD's bridge has been rewritten from scratch and this is a supported setup, I figured others must be leveraging similar architectures to blend legacy gigabit gear with newer multi-gig hardware. Appreciate the insights from everyone—it's always great to pressure-test network designs!
Quote from: Patrick M. Hausen on September 29, 2026, 01:20:12 PMThese definitions match the interface configuration. I am out of ideas ... er ... one moment ...
Did you check this option?
(https://forum.opnsense.org/index.php?action=dlattach;attach=58356;image)
If you did not that would perfectly explain your observed symptoms.
Quote from: Patrick M. Hausen on September 29, 2026, 01:20:12 PMThese definitions match the interface configuration. I am out of ideas ... er ... one moment ...
Did you check this option?
(https://forum.opnsense.org/index.php?action=dlattach;attach=58356;image)
If you did not that would perfectly explain your observed symptoms.
ISC creates the necessary firewall rules for DHCP to work by default, if I remember correctly. For Kea that's optional, because some users complained that OPNsense should not create any rules automatically giving full control to the admin.
ISC creates the necessary firewall rules for DHCP to work by default, if I remember correctly. For Kea that's optional, because some users complained that OPNsense should not create any rules automatically giving full control to the admin.
Hi @Patrick,
Do you have any further feedback on this? Since I've already exhausted everything possible on the Kea DHCP side, I wanted to ask: could this potentially be a bug?
If so, is there anything else I can do to help get this reviewed before the next release to confirm whether it's an issue?
Cheers
What kind of feedback do you expect? You have not yet answered my question. Is "Firewall rules" active in your setup?
Yes it is, that mode was enabled by default.
I had the same setting as your picture.
Turning it off then on again made no difference
regards