Messages

Upgrade okay, Init7 25gbps on iperf3 with E810+ddp works good.

XGS-PON is becoming available from my service providers and I have started to look for hardware to facilitate migration to these services.

While I can find hardware with combinations of 2.5Gbit/s copper with SPF+, I am yet to find much in the SOHO market with 10Gbit/s copper with SPF+.

Does anyone have any recommendations for hardware that will run OPNsense?

I run 25gbps internet on MS-A2 9955hx. In past used 14700, non-k version.

Important consideration is the network card that you will use. E810 or Connect4-x Lx both work great for me.

I get 25gbps routing performance out of OpenSense easily on 8 cores

I want to say I might have same problem, but I'm not sure at this stage.

I do see same RSS errors in dmesg.

Recently my internet speed dropped from 25gbp/s to 8-9. I do run Intel E810-XXV dual port adapter on Opnsense 25.1.10
I used to get 25gbp/s throughput via Opnsense as verified via iperf3 tests (both ipv4 and ipv6) from LAN client to Internet iperf3 server.

What puzzles me, is that I easily get 25gbp/s in iperf3 by running it on LAN interface of OPnsense from a LAN client.
Also, when I look at top during the above test, I see all CPUs busy doing something :

Code Select Expand

last pid: 12095;  load averages:  0.95,  0.45,  0.28                                                                                                                                      up 0+01:48:02  21:13:50
83 processes:  1 running, 82 sleeping
CPU 0:  0.0% user,  0.0% nice, 56.3% system,  0.4% interrupt, 43.4% idle
CPU 1:  0.4% user,  0.0% nice, 23.0% system,  2.7% interrupt, 73.8% idle
CPU 2:  3.5% user,  0.0% nice,  2.7% system, 16.0% interrupt, 77.7% idle
CPU 3:  0.4% user,  0.0% nice, 26.2% system,  1.6% interrupt, 71.9% idle
CPU 4:  0.4% user,  0.0% nice, 68.4% system,  5.1% interrupt, 26.2% idle
CPU 5:  0.0% user,  0.0% nice, 57.4% system,  1.2% interrupt, 41.4% idle
CPU 6:  0.4% user,  0.0% nice, 57.4% system,  0.4% interrupt, 41.8% idle
CPU 7:  0.0% user,  0.0% nice, 69.5% system,  0.0% interrupt, 30.5% idle
Mem: 142M Active, 313M Inact, 763M Wired, 305M Buf, 6643M Free


So to recap :

LAN client -> ice1 LAN Opnsense (iperf3 -s) -> 25gpb/s easily
LAN client -> ice1 -> ice0 (WAN) -> Internet iperf3 -> Dropped from 24-25 to 8-9


At this stage, I'm not sure if this is Opnsense or my ISP...

I read this thread and I still don't understand few things.


I use Unbound as DNS and don't want to change to Dnsmasq.

I'm all in favor to drop ISC DHCP and migrate to Kea but I need Router Advertisement support for IPV6.

What are my options ?

That is easy, you use:
- Services/Unbound DNS
- Services/Kea DHCPv4
- Services/Router Advertisements

Apologies, I also use DHCPv6 and RA runs in Assisted mode today.

It's a standard setup, I get IPV6 prefix from ISP and I use DHCPv6 to assign IPs from one of the subnets.

I read this thread and I still don't understand few things.


I use Unbound as DNS and don't want to change to Dnsmasq.

I'm all in favor to drop ISC DHCP and migrate to Kea but I need Router Advertisement support for IPV6.

What are my options ?

I added few backends to Upstream, then used them in SNI based routing. After I finished testing, removed SNI based routing but enable to remove unused Upstream.

Just getting "Item in used by" message

Code Select Expand

Nginx - {Nginx.sni_hostname_upstream_map_item.5dbfd85e-f1aa-43ca-bfa2-313a684a199c}
Item is not referenced anywhere 100%, nothing at all in nginx.conf

Tried to manually grep files in /etc or /usr/local/etc - no matches at all.

Nginx restart or reboot didn't help at all




Any ideas how to fix this ?

i7-14700
Proxmox
Opnsense 25.1.1
Mellanox ConnectX-4 Lx
1 port passthrough WAN
1 port bridged LAN into opnsense and TrueNAS


speedtest from TrueNAS

Code Select Expand

# bin/speedtest -s 43030

   Speedtest by Ookla

      Server: Init7 AG - Winterthur (id: 43030)
         ISP: Init7
Idle Latency:     1.44 ms   (jitter: 0.08ms, low: 1.42ms, high: 1.70ms)
    Download: 23453.57 Mbps (data used: 23.0 GB)
                  3.30 ms   (jitter: 3.67ms, low: 1.13ms, high: 26.84ms)
      Upload: 22000.78 Mbps (data used: 22.7 GB)
                  1.25 ms   (jitter: 0.11ms, low: 1.08ms, high: 1.94ms)
 Packet Loss:     0.0%

https://www.speedtest.net/result/c/db97cbad-a4d3-4d23-af27-980535ffbe23

To finish this up ...

The Intel E810-XXVDA2 needs some developer love to work with OPNsense, I change to Broadcom P225P and this card works out of the box without compiling  kernel or ports.

Thanks Netnut for your help !

Did you try 25.1 ?

ice_ddp driver has been upgraded to latest version

So I'm on 25.1 and I get same error for ice_ddp.ko in dmesg

Code Select Expand

# dmesg | grep ice_ddp
ice0: The DDP package module (ice_ddp) failed to load or could not be found. Entering Safe Mode.
ice0: The DDP package module cannot be automatically loaded while booting. You may want to specify ice_ddp_load="YES" in your loader.conf
ice1: The DDP package module (ice_ddp) failed to load or could not be found. Entering Safe Mode.
ice1: The DDP package module cannot be automatically loaded while booting. You may want to specify ice_ddp_load="YES" in your loader.conf

EDIT: Actually never mind, I added both of these :
ice_ddp_load = YES and if_ice_load = YES in Tunnables

and by the looks of it, everything works now :

Code Select Expand

ice0: <Intel(R) Ethernet Network Adapter E810-XXV-2 - 1.43.2-k> mem 0x380000000000-0x380001ffffff,0x380002000000-0x38000200ffff irq 16 at device 0.0 on pci1
ice0: Loading the iflib ice driver
ice0: DDP package already present on device: ICE OS Default Package version 1.3.41.0, track id 0xc0000001.
ice0: fw 7.3.4 api 1.7 nvm 4.30 etid 8001b891 netlist 4.2.5000-1.14.0.2b9b23c0 oem 1.3415.0
ice0: Using 8 Tx and Rx queues
ice0: Reserving 8 MSI-X interrupts for iRDMA
ice0: Using MSI-X interrupts with 17 vectors
ice0: Using 1024 TX descriptors and 1024 RX descriptors
ice0: Ethernet address: 50:7c:6f:79:ca:e8
ice0: ice_add_rss_cfg on VSI 0 could not configure every requested hash type
ice0: PCI Express Bus: Speed 16.0GT/s Width x8
ice0: Firmware LLDP agent disabled
ice0: Link is up, 10 Gbps Full Duplex, Requested FEC: None, Negotiated FEC: None, Autoneg: False, Flow Control: None
ice0: link state changed to UP
ice0: netmap queues/slots: TX 8/1024, RX 8/1024


Does 25.1 now fully support ice_ddp driver ?