Hi all,
I have an OpenWrt One (bought Feb 2025, original 12V PD supply) with intermittent packet loss on the WAN, and I’ve run out of things to rule out on my side. I’d appreciate any ideas, or confirmation if this looks like a hardware fault.
Setup: cable ISP (Tigo, Colombia), modem in bridge mode, OpenWrt One as main router (AdGuard Home + Tailscale).
Symptom: ping to 8.8.8.8 from the router loses 7–42% of packets in 300–1000 packet runs. It comes and goes on its own (clean periods of minutes to a few hours). Link stays up at 1000/Full, no link flaps in logs, rx_crc_errors 0, temperature ~64 °C.
What I tested:
- OpenWrt 25.12.6 and 24.10.8: both show the loss, including a factory-default config with Wi-Fi off and no LAN devices connected.
- WAN on the 2.5G port (EN8811H) and on the 1G port (eth1, reassigned as WAN): loss on both.
- Limiting the 2.5G port to advertise 1000baseT only: no change.
- 3 cables, 2 modem ports, 2 MAC addresses, 2 power supplies (original 12V PD and an Anker 140W PD): no change.
- Laptop connected directly to the second modem port at the same time the router was losing 19–34%: 0% loss (3 runs of 300).
- TP-Link Archer A6 v2 (OpenWrt 25.12.6) on the same modem and cable: 0 lost out of 1000.
- After a modem reboot, the One dropped to 8/1000 lost. I’m now logging loss every 5 minutes to see if it comes back.
Questions:
- Has anyone seen similar behavior on the OpenWrt One, on either port?
- Is there any diagnostic you’d suggest (counters, debug, PHY/MAC registers) that would show where the packets are lost?
- Could this be a hardware issue with the MT7981B Ethernet, since both ports are affected?
Port counters on eth1 (current WAN) are clean:
rx_fcs_errors: 0
rx_short_errors: 0
rx_long_errors: 0
rx_checksum_errors: 1
rx_xdp_drop: 0
rx_xdp_tx_errors: 0
tx_xdp_xmit_errors: 0
So whatever is being lost doesn’t show up as RX errors at the MAC.
I can post full command outputs. Thanks!