Hi,
A switch (Netgear GS108T v3, 21.01.1) shows warnings in the log and drops packets aimed towards the device. Traffic entering a port and leaving another port is unaffected/fine:
Tue Nov 9 17:29:43 2021 kern.warn kernel: [88189.851532] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a700003c, cur a700004c
Tue Nov 9 17:29:43 2021 kern.warn kernel: [88189.866290] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a7000040, cur a700004c
Tue Nov 9 17:29:43 2021 kern.warn kernel: [88189.889970] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a7000044, cur a700004c
Tue Nov 9 17:29:43 2021 kern.warn kernel: [88189.969803] rtl838x-eth bb00a300.ethernet eth0: Ring
...
Tue Nov 9 17:30:28 2021 kern.info kernel: [88234.247018] RX buffer overrun: status 101, mask: ffeff
Tue Nov 9 17:30:28 2021 kern.info kernel: [88234.252886] RX buffer overrun: status 1, mask: ffeff
Tue Nov 9 17:30:28 2021 kern.info kernel: [88234.258510] RX buffer overrun: status 1, mask: ffeff
The error just occurs after a while:
The issue can be triggered by running iperf3:
iperf3 -c 192.168.xx.254 -t0
Connecting to host 192.168.xx.254, port 5201
[ 5] local 192.168.xx.13 port 44924 connected to 192.168.xx.254 port 5201
[ ID] Interval Transfer Bitrate Retr Cwnd
[ 5] 0.00-1.00 sec 4.69 MBytes 39.2 Mbits/sec 0 233 KBytes
[ 5] 1.00-2.05 sec 2.75 MBytes 22.1 Mbits/sec 0 369 KBytes
[ 5] 2.05-3.26 sec 3.57 MBytes 24.7 Mbits/sec 0 533 KBytes
[ 5] 3.26-4.03 sec 1.12 MBytes 12.3 Mbits/sec 1 551 KBytes
[ 5] 4.03-5.00 sec 0.00 Bytes 0.00 bits/sec 1 551 KBytes
[ 5] 5.00-6.00 sec 1.11 MBytes 9.34 Mbits/sec 0 420 KBytes
[ 5] 6.00-7.00 sec 1.37 MBytes 11.5 Mbits/sec 0 451 KBytes
...
[ 5] 99.00-100.00 sec 0.00 Bytes 0.00 bits/sec 0 1.72 MBytes
[ 5] 100.00-101.00 sec 0.00 Bytes 0.00 bits/sec 0 1.72 MBytes
^C[ 5] 101.00-101.93 sec 0.00 Bytes 0.00 bits/sec 0 1.72 MBytes
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bitrate Retr
[ 5] 0.00-101.93 sec 66.2 MBytes 5.45 Mbits/sec 6 sender
[ 5] 0.00-101.93 sec 0.00 Bytes 0.00 bits/sec receiver
iperf3: interrupt - the client has terminated
On a separate console the log shows this while iperf3 is active:
Tue Nov 9 17:49:50 2021 kern.info kernel: [ 903.051213] RX buffer overrun: status 101, mask: ffeff
Tue Nov 9 17:49:50 2021 kern.info kernel: [ 903.057182] RX buffer overrun: status 101, mask: ffeff
Tue Nov 9 17:49:50 2021 kern.info kernel: [ 903.063026] RX buffer overrun: status 1, mask: ffeff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 903.937196] RX buffer overrun: status 101, mask: ffcff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 903.943113] RX buffer overrun: status 1, mask: ffcff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 903.948754] RX buffer overrun: status 1, mask: ffcff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 903.954392] RX buffer overrun: status 101, mask: ffcff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 903.960362] RX buffer overrun: status 101, mask: ffcff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 903.966204] RX buffer overrun: status 1, mask: ffcff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 904.056387] RX buffer overrun: status 101, mask: ffeff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 904.062297] RX buffer overrun: status 1, mask: ffeff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 904.067934] RX buffer overrun: status 1, mask: ffeff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 904.073570] RX buffer overrun: status 101, mask: ffeff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 904.079537] RX buffer overrun: status 101, mask: ffeff
Tue Nov 9 17:49:51 2021 kern.info kernel: [ 904.085384] RX buffer overrun: status 1, mask: ffeff
After terminating iperf3 the message changes and it remains as follows even though the high load situation is now gone:
Tue Nov 9 17:49:59 2021 kern.warn kernel: [ 911.986238] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000c4, cur a70000c8
Tue Nov 9 17:49:59 2021 kern.warn kernel: [ 911.996379] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000c4, cur a70000c8
Tue Nov 9 17:50:00 2021 kern.warn kernel: [ 913.020755] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000cc, cur a70000d8
Tue Nov 9 17:50:00 2021 kern.warn kernel: [ 913.030895] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000cc, cur a70000d8
Tue Nov 9 17:50:00 2021 kern.warn kernel: [ 913.242054] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000d0, cur a70000d8
Tue Nov 9 17:50:00 2021 kern.warn kernel: [ 913.258000] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000d4, cur a70000d8
Tue Nov 9 17:50:01 2021 kern.warn kernel: [ 914.217759] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000e8, cur a70000ec
Tue Nov 9 17:50:01 2021 kern.warn kernel: [ 914.227900] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000e8, cur a70000ec
Tue Nov 9 17:50:50 2021 kern.warn kernel: [ 963.623829] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000c4, cur a70000c8
Tue Nov 9 17:50:50 2021 kern.warn kernel: [ 963.633968] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000c4, cur a70000c8
Tue Nov 9 17:50:50 2021 kern.warn kernel: [ 963.830442] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000cc, cur a70000d8
Tue Nov 9 17:50:50 2021 kern.warn kernel: [ 963.840583] rtl838x-eth bb00a300.ethernet eth0: Ring contention: r: 0, last a70000cc, cur a70000d8
Traffic passing the switch from one port to another is not affected, it seems to pause traffic to eth0 for brief moments. To observe that a ping -A 192.168.xx.254 can be used.
Is this a known issue? I do not mind that the device is easily overwhelmed, I am just surprised that the device enters a state where it is loosing packets even if high traffic stops and remains in this poor state.