Unable to Configure ECN on S5248F-ON – ecnconfig Fails and No ECN Marking on Packets

[erfiboy]

Description

Hello,

I am currently facing an issue configuring ECN on an S5248F-ON switch running SONiC. The default QoS template (5248-on) does not include WRED, and I am unable to configure it manually.
When I replace the default template with the generic one (which includes WRED configurations), I can see the ECN settings. However, when I try to apply ECN to the queues, I encounter an error. Am I doing something wrong? I even dumped the packets, and none of them had ECN markings.

Steps to reproduce the issue:

1. Install the SONiC.202405.704700-29cbc5423 on the Dell S5248-ON switch.
2. Run the command ecnconfig -l. You see nothing, and there is no profile available to set the ECN parameters.
3. Replace the json file at /usr/share/sonic/device/x86_64-dellemc_s5248f_c3538-r0/DellEMC-S5248f-P-25G/qos.json.j2 with the generic one from link or use the one available on the device.
4. Apply changes by running sudo config qos reload. Now you can set ECN parameters, but the command ecnconfig -q 3,4 fails.

Describe the results you received:

admin@switch:~$ ecnconfig -l
Profile: AZURE_LOSSLESS
-----------------------  -------
ecn                      ecn_all
green_drop_probability   5
green_max_threshold      2097152
green_min_threshold      1048576
red_drop_probability     5
red_max_threshold        2097152
red_min_threshold        1048576
wred_green_enable        true
wred_red_enable          true
wred_yellow_enable       true
yellow_drop_probability  5
yellow_max_threshold     2097152
yellow_min_threshold     1048576
-----------------------  -------

admin@switch:~$ sudo ecnconfig -p AZURE_LOSSLESS -gmin 100000 -gmax 250000 -ymin 50000 -ymax 200000 -rmin 25000 -rmax 150000

admin@switch:~$ ecnconfig -l
Profile: AZURE_LOSSLESS
-----------------------  -------
ecn                      ecn_all
green_drop_probability   5
green_max_threshold      250000
green_min_threshold      100000
red_drop_probability     5
red_max_threshold        150000
red_min_threshold        25000
wred_green_enable        true
wred_red_enable          true
wred_yellow_enable       true
yellow_drop_probability  5
yellow_max_threshold     200000
yellow_min_threshold     50000
-----------------------  -------

admin@switch:~$ ecnconfig -q 3
Exception caught:  No active ports detected in table 'DEVICE_NEIGHBOR'

Output of show version:

SONiC Software Version: SONiC.202405.704700-29cbc5423
SONiC OS Version: 12
Distribution: Debian 12.6
Kernel: 6.1.0-22-2-amd64
Build commit: 29cbc5423
Build date: Mon Nov 25 13:46:52 UTC 2024
Built by: azureuser@5348dadec000001

Platform: x86_64-dellemc_s5248f_c3538-r0
HwSKU: DellEMC-S5248f-P-25G
ASIC: broadcom
ASIC Count: 1
Serial Number: CJ9RY03
Model Number: 0F16J9
Hardware Revision: N/A
Uptime: 21:20:41 up 8 days, 22:10,  1 user,  load average: 0.93, 0.77, 0.61
Date: Fri 28 Feb 2025 21:20:41

Docker images:
REPOSITORY                    TAG                       IMAGE ID       SIZE
docker-syncd-brcm             202405.704700-29cbc5423   de02c3c90ae5   742MB
docker-syncd-brcm             latest                    de02c3c90ae5   742MB
docker-gbsyncd-broncos        202405.704700-29cbc5423   89da1429894b   352MB
docker-gbsyncd-broncos        latest                    89da1429894b   352MB
docker-gbsyncd-credo          202405.704700-29cbc5423   722263daa960   326MB
docker-gbsyncd-credo          latest                    722263daa960   326MB
docker-orchagent              202405.704700-29cbc5423   59fce64fe2ca   356MB
docker-orchagent              latest                    59fce64fe2ca   356MB
docker-sflow                  202405.704700-29cbc5423   156db8a52858   344MB
docker-sflow                  latest                    156db8a52858   344MB
docker-nat                    202405.704700-29cbc5423   994472145d1b   345MB
docker-nat                    latest                    994472145d1b   345MB
docker-fpm-frr                202405.704700-29cbc5423   535d919af5d9   375MB
docker-fpm-frr                latest                    535d919af5d9   375MB
docker-dhcp-relay             latest                    8cc7f6110640   324MB
docker-macsec                 latest                    d176c0c23727   345MB
docker-platform-monitor       202405.704700-29cbc5423   0e32c23e19d7   434MB
docker-platform-monitor       latest                    0e32c23e19d7   434MB
docker-teamd                  202405.704700-29cbc5423   0bb6d977cf02   343MB
docker-teamd                  latest                    0bb6d977cf02   343MB
docker-snmp                   202405.704700-29cbc5423   60e192b6274a   354MB
docker-snmp                   latest                    60e192b6274a   354MB
docker-router-advertiser      202405.704700-29cbc5423   5e8b1e3d546a   315MB
docker-router-advertiser      latest                    5e8b1e3d546a   315MB
docker-eventd                 202405.704700-29cbc5423   af91ec0250f6   314MB
docker-eventd                 latest                    af91ec0250f6   314MB
docker-lldp                   202405.704700-29cbc5423   1394557ec7ed   360MB
docker-lldp                   latest                    1394557ec7ed   360MB
docker-mux                    202405.704700-29cbc5423   c76677082f04   366MB
docker-mux                    latest                    c76677082f04   366MB
docker-sonic-gnmi             202405.704700-29cbc5423   daac8816515f   399MB
docker-sonic-gnmi             latest                    daac8816515f   399MB
docker-database               202405.704700-29cbc5423   32918209c5a7   323MB
docker-database               latest                    32918209c5a7   323MB
docker-sonic-mgmt-framework   202405.704700-29cbc5423   c0532c6aea4b   401MB
docker-sonic-mgmt-framework   latest                    c0532c6aea4b   401MB

Output of show techsupport:

The size of the tech support file was larger than 25MB, so I uploaded it to a storage service. Here is the link

Additional information you deem important (e.g. issue happens only occasionally):

admin@switch:~$ show platform firmware status

Chassis    Module    Component     Version      Description
---------  --------  ------------  -----------  ---------------------------------------------------------------------------------------------------------
S5248F-ON  N/A       BIOS          3.40.0.9-17  Performs initialization of hardware components during booting
                     FPGA          3.4          Used for managing the system LEDs
                     BMC           1.8          Platform management controller for on-board temperature monitoring, in-chassis power, Fan and LED control
                     System CPLD   0.9          Used for managing the CPU power sequence and CPU states
                     Slave CPLD 1  1.0          Used for managing SFP28/QSFP28 port transceivers (SFP28 1-24, QSFP28 1-4)
                     Slave CPLD 2  1.0          Used for managing SFP28/QSFP28 port transceivers (SFP28 25-48, QSFP28 5-8)

[erfiboy]

@jeff-yin @anilkpan
Could you please take a look at this issue? Let me know if you need any more details. Thanks!

[jeff-yin]

I've asked some Dell team members to help take a look, although it's not clear yet whether this is a platform-specific issue.

[erfiboy]

@jeff-yin
Hi Jeff,

Thanks for your response. WRED is not enabled in the QoS configuration of the Dell switch. However, it should be, as this switch is supposed to support ECN marking. Since ECN should be part of the default QoS configuration, I think this issue is related to Dell.

However, the template includes a WRED configuration: Template QOS configuration

Let me know your thoughts.

[AshokDaparthi]

@erfiboy - i hope you did this and tried to generate configs.

in the file device/dell/x86_64-dellemc_s5248f_c3538-r0/DellEMC-S5248f-P-25G-DPB/qos.json.j2, replace everything with below line.
{%- include 'qos_config.j2' %}

Even aftet doing this if it not working. please send me below output. /tmp/cfg_qos.json

/usr/local/bin/sonic-cfggen -d -t /usr/share/sonic/device/x86_64-dellemc_s5248f_c3538-r0/DellEMC-S5248f-P-25G/qos.json.j2,/tmp/cfg_qos.json -y /etc/sonic/sonic_version.yml

[erfiboy]

@AshokDaparthi, Thanks for your response.

I replaced {%- include 'qos_config.j2' %} with the entire content of the file and reloaded the QoS configuration. As a result, I can see the ECN configurations as if I had copied the generic configuration into this file and reloaded the QoS configuration.

However, I still encounter an error when running the following command:

ecnconfig -q 3
Exception caught: No active ports detected in table 'DEVICE_NEIGHBOR'

Here is the output you mentioned.

Although I doubt the problem is resolved, I have connected two NICs to the switch, with the link operating at 25 Gbps. Each of these switches is connected to a separate VRF, and the VRFs are linked via a 10 Gbps connection. I am sending traffic using RDMA.

When checking the drop counters, I see a large number of packets being dropped at the interface where the NIC is connected to the switch. However, I don’t see any ECN-marked packets.

[arlakshm]

@jeff-yin, assigned this issue to you as this dell platform. Please reassign if this is not platform issue