I have just changed my Netgear GS108e (8x1Gb) switch with a GM110mx (8x1Gb + 2x10Gb), and noticed that I have now some rx_csum_offload_errors and rx_length_errors errors on my Intel x550T2 NIC stats of my NAS. Those errors don't increase very often, but more quickly under heavy transfer loads.
rx_errors: 692
tx_errors: 0
rx_over_errors: 0
rx_crc_errors: 0
rx_frame_errors: 0
rx_fifo_errors: 0
rx_missed_errors: 0
tx_aborted_errors: 0
tx_carrier_errors: 0
tx_fifo_errors: 0
tx_heartbeat_errors: 0
rx_length_errors: 135
rx_long_length_errors: 0
rx_short_length_errors: 0
rx_csum_offload_errors: 557
fcoe_last_errors: 0
The rx-errors stats seem to be the sum of rx_length_errors and rx_csum_offload_errors. There also seems to be some checksum and length errors.
The NAS and my PC are both on the 2x10Gb/S switch port, the ISP box is plugged on 1x1Gb/s port. You can see my NAS configuration here: Intel x550 NIC rx_dropped packet on Debian 10
From what I have seen, the counter does not increase when I copy some files from my PC to my NAS, and reverse, at 2,5Gb/s speed. But it increases when I download some files directly from the internet with a torrent app or a wget command (at 1Gb/s speed).
wget -O /dev/null https://cdimage.debian.org/debian-cd/current/amd64/iso-dvd/debian-11.5.0-amd64-DVD-1.iso)
The MTU of the ISP box and the switch are both set to 1500 I think, but cannot confirm that as this is an unmanaged switch.
My ISP box stats are clear but not very detailed :(
tx_good_bytes = 464828503658
tx_collisions = 0
tx_unicast_pkts = 398259904
tx_multicast_pkts = 325088
tx_broadcast_pkts = 14177
rx_good_bytes = 315076986100
rx_fcs_errors = 0
rx_unicast_pkts = 331564110
So, is there a way to know the origin (the device) of those errors ?
