Hello OPNsense Community,
I am struggling with a persistent WAN connection issue and have reached the end of my troubleshooting knowledge. I would be very grateful for any insights the community can offer.
The Problem:
Starting around September 15th, my internet connection drops intermittently, typically every 45-60 minutes, but the timing is not exact. These drops are severe enough to disconnect me from video meetings and disrupt other real-time traffic. The issue does not seem to be tied to the DHCP lease renewal anymore.
Hardware & Software:
OPNsense Version: OPNsense 25.7.5-amd64 (Virtualized through Proxmox)
Hardware: Topton mini PC with quad Intel i226-V 2.5G NICs (using the igc driver).
WAN Interface: igc1, manually set to 2500baseT. (Provided to the OPNsense VM via PCI Passthrough)
Upstream Device: AT&T BGW320-505 (In passthrough mode)
Detailed Troubleshooting Steps Already Taken:
I have performed extensive troubleshooting to isolate the cause, including:
DHCP Lease: Initially, I found a 5-minute DHCP lease and changed it to 1 hour. This seemed to fix the problem for about 8 days, but the intermittent drops returned. I have since changed the lease time to 24 hours, but this has had no noticeable effect on the frequency of the drops. The drops no longer correlate with the DHCP renewal in the logs.
Intel i226-V Firmware: I have updated the firmware on all i226-V NICs to the latest available version (2.32). This did not change the behavior.
Kernel Tunables for i226-V: I have set the recommended kernel tunables for igc driver stability.
hw.igc.eee_setting = 1 (Disabled Energy Efficient Ethernet)
hw.igc.rx_abs_int_delay = 0
hw.igc.tx_abs_int_delay = 0
hw.pci.enable_aspm = 0
Hardware Offloading: All hardware offloading (Checksum, TSO, LRO) was already disabled on the WAN interface from the initial setup of the firewall.
Gateway Monitoring: Gateway logs from earlier in the troubleshooting process did show that the drops were caused by packet loss crossing the "down" threshold (Loss: 22.0 %). I have since made the monitoring more tolerant, but the underlying packet loss issue persists and still kills active connections.
My Core Question:
Given that the issue is no longer tied to DHCP renewals and that I've already updated the NIC firmware and applied all the common igc driver tunables, what could be the source of this intermittent link instability?
It feels like a hardware compatibility or link-layer issue between the i226-V NIC and my ISP's gateway that isn't being solved by the standard software/driver fixes. Are there any other advanced tunables, diagnostics, or known issues with this hardware combination that I should investigate?
Thank you for your time and assistance. I've attached a quality graph that shows the intermittent nature of the packet loss.
(https://i.imgur.com/LGstnuc.png)
(https://imgur.com/csPDYGE.png)
Read a lot of this thread
https://forum.opnsense.org/index.php?topic=48695.0
I did go through that thread in my initial troubleshooting. It was the reason I upgraded my NIC fw from 2.17 to 2.32 in hopes that it would fix this issue. Is there something specific I missed other than the firmware upgrades and tunables?
This is the output from nvmupdate:
[00:006:16:00]: Intel(R) Ethernet Controller I226-V
Vendor : 8086
Device : 125C
Subvendor : 8086
Subdevice : 0000
Revision : 4
LAN MAC : 00D04C100736
Alt MAC : 000000000000
SAN MAC : 000000000000
ETrackId : 80000425
SerialNumber : N/A
NVM Version : 2.50(2.32)
PBA : G23456-000
VPD status : Not set
VPD size : 0
NVM update : No config file entry
checksum : Valid
[00:006:17:00]: Intel(R) Ethernet Controller I226-V
Vendor : 8086
Device : 125C
Subvendor : 8086
Subdevice : 0000
Revision : 4
LAN MAC : 00D04C100737
Alt MAC : 000000000000
SAN MAC : 000000000000
ETrackId : 80000425
SerialNumber : N/A
NVM Version : 2.50(2.32)
PBA : G23456-000
VPD status : Not set
VPD size : 0
NVM update : No config file entry
checksum : Valid
I've gone through and disabled ASPM in the BIOS of the minipc, I'll monitor for any changes since it's applied to the Proxmox host now.
That didn't seem to help either. Still seeing the intermittent losses in the Quality graph and I get dropped from matches while playing games.
Does pciconf for igc show aspm as disabled? You made cfg settings, now you need to verify.
The system log files should be logging interface state changes, what do the logs say?