NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
M4300-28G-POE+: LAG + Port Mirroring results in high number of Transmit Packet Discards
Hello,
I have an issue with a M4300-28G-POE+ model. The setup I have is quite simple:
* 2x1 Gbit links are used which are batched into a LAG
* The LAG is mirrored to a probe port.
* Jumbo packets are configured.
When I generate network traffic on the two links, the 10Gbit uplink receives each of the packets of the LAG of course, but when I am listening on the probe port, ~40% of the packets are dropped. In the Detailed Port Statistics I can see the following:
Total Packets Transmitted Successfully: 161364755
Transmit Packets Discarded: 73094691
This example was taken from a scenario when the overall network traffic on the LAG was ~170Mbit/s and on the probe port I measured ~100Mbit/s
I've tried to change the configuration based on hints in the Netgear Community threads (like this, since the observation is quite similar: https://community.netgear.com/t5/Smart-Plus-Click-Switches/GS724T-High-number-of-Transmit-Packet-Errors/m-p/393029#M257). When I am not using Jumbo packets but LAG, I observe the same issue.
I've updated the firmware to 12.0.2.15 but the issue still occurs on my test environment.
Are there any suggestions how to deal with this problem?
Best,
I.
5 Replies
Thank you for that. It shouldn't have any discarded transmit packets. Kindly give us more information about this.
- May you please give us an idea what are the devices that you are using on your test environment?
- Purpose in setting up LAGs
- Device that will be connected to LAG
We may need to further check this using SYSLOG server.
I look forward for your reply.
Regards,
Dear JohnC_V,
First of all, the devices in the environment:
- I am using a GigE Vision compliant industrial camera (2x1Gbit links) which data stream shall be mirrored to a probe port for further data processing.
- The uplink is connected to a workstation which is equipped with 10Gbit Ethernet card.
- The probe port is connected to a desktop computer which is equipped with 1Gbit Ethernet card.
- The cabling is established with CAT 6 cables.
- The packet size in this topology is mutable (but preferably 9000 bytes Jumbo packets shall be used). I performed measurements with packet sizes from 1000 bytes up to 9000 bytes, but I observed the same loss on the probe port.
Purpose in setting up LAGs
- The industrial cameras are equipped with 2x1Gbit ports (with PoE) and the two links transfers the raw camera image frames over GVSP protocol and the links shall be in the same aggregation group. Without a static mode LAG the camera driver can't access the camera.
- Device that will be connected to LAG
- The LAG consists of the 2x1Gbit ports of the camera. The 2x1Gbit links are the trunk members and the LAG settings are the following:
Admin mode: Enabled
Hash Mode: 3 Src/Dest MAC, VLAN, EType, incoming port
STP Mode: Enable
Static Mode: Enable
Link Trap: Enable
Configured ports: port1, port2
Active ports: port1, port2
LAG state: Up
Local Preference Mode: DisablePlease let me know if you need further information.
Best Regards,
I.
Thank you for that information. We may need a syslogs on this so that we can compare it to where is the discarded packets coming from. The best way that we can do in here is that you need to open a chat / case online with NETGEAR support and please let them know the status of the switch so that they can review your setup and identify why there is some discarded packets on LAG.
Regards,
Hi foobar12345,
It seems that you were having problem with LAGs on your switch. May you be able to try resetting it to factory default and then just enable LAG on the same port without changing any MTU or other settings. Then try to connect it back again and see if it will show the same issue.
Welcome to the community!
Regards,
Dear JohnCarloV,
I've reset to factory default settings as suggested. The MTU is set now back to 1500 bytes. I've re-established the LAG (without any further changes to the MTU) with the 2x1Gbit ports and mirrored the LAG to a probe port. The observed result is the same with regards to the bandwidth and Discarded Transmit Packets. I've replayed the scenario (with the expected ~170Mbit/s) BUT with 1500 bytes packet size, but on the probe port I can still measure the ~100Mbit/s network load only.
Are there any assumptions or hints what else could I try?
Best Regards,
I.
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
