X710-DA4 packet drops

[MCecc1]

Hi,

we are running some test on the X710 cards and packets are dropping big time. The test is pretty standard, using a physical traffic generator the traffic flows in port 1 and out from port 2 of the same. At first virtual machine with SR-IOV VF were used (up to 4 Gbps of traffic everything was running fine), but since the traffic started dropping the virtualization layer was taken out of the scene in order to troubleshoot the issue. Beyond 4 Gbps with a packet size of 1500b things start going bad, while using a jumbo frame the throughput is much better.

OS in use is Ubuntu 18.04.

Ethernet controller: Intel Corporation Ethernet Controller X710 for 10GbE SFP+ (rev 01)   Subsystem: Cisco Systems Inc Ethernet Converged NIC X710-DA   Flags: bus master, fast devsel, latency 0, IRQ 78, NUMA node 0   Memory at c3800000 (64-bit, prefetchable) [size=8M]   Memory at c5808000 (64-bit, prefetchable) [size=32K]   Expansion ROM at c5d80000 [disabled] [size=512K]   Capabilities: [40] Power Management version 3   Capabilities: [50] MSI: Enable- Count=1/1 Maskable+ 64bit+   Capabilities: [70] MSI-X: Enable+ Count=129 Masked-   Capabilities: [a0] Express Endpoint, MSI 00   Capabilities: [e0] Vital Product Data   Capabilities: [100] Advanced Error Reporting   Capabilities: [140] Device Serial Number e8-19-c2-ff-ff-fe-fd-3c   Capabilities: [150] Alternative Routing-ID Interpretation (ARI)   Capabilities: [160] Single Root I/O Virtualization (SR-IOV)   Capabilities: [1a0] Transaction Processing Hints   Capabilities: [1b0] Access Control Services   Kernel driver in use: i40e   Kernel modules: i40edriver: i40e version: 2.7.29 firmware-version: 5.05 0x80002b50 0.385.33 expansion-rom-version: bus-info: 0000:5e:00.2 supports-statistics: yes supports-test: yes supports-eeprom-access: yes supports-register-dump: yes supports-priv-flags: yesArchitecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 40 On-line CPU(s) list: 0-39 Thread(s) per core: 1 Core(s) per socket: 20 Socket(s): 2 NUMA node(s): 2 Vendor ID: GenuineIntel CPU family: 6 Model: 85 Model name: Intel(R) Xeon(R) Gold 6138 CPU @ 2.00GHz Stepping: 4 CPU MHz: 1000.000 CPU max MHz: 2000.0000 CPU min MHz: 1000.0000 BogoMIPS: 4000.00 Virtualization: VT-x L1d cache: 32K L1i cache: 32K L2 cache: 1024K L3 cache: 28160K NUMA node0 CPU(s): 0-19 NUMA node1 CPU(s): 20-39    

[Mike_Intel]

Hello MCecc1, Thank you for posting in Intel Ethernet Communities. Please provide further information below for us check the issue: 1) What is the tool you used to see the packet loss? 2) Run the command ethtool –i to show the driver, firmware version and other information If you have questions, please let us know. Best regards, Michael L. Intel Customer Support Under Contract to Intel Corporation

[MCecc1]

​Hi Mike,

thanks for  your answer.

The drops are observed using standard Linux commands ( ie netstat -i ) and below you can find again the firmware driver info.

driver: i40e version: 2.7.29 firmware-version: 5.05 0x80002b50 0.385.33 expansion-rom-version: bus-info: 0000:5e:00.2 supports-statistics: yes supports-test: yes supports-eeprom-access: yes supports-register-dump: yes supports-priv-flags: yes

[Mike_Intel]

Hello MCecc1, We have UDP issue stated in readme file but is applicable to ixgbe driver, sharing with you as reference. UDP Stress Test Dropped Packet Issue ------------------------------------ Under small packet UDP stress with the ixgbe driver, the system may drop UDP packets due to socket buffers being full. Setting the driver Intel Ethernet Flow Control variables to the minimum may resolve the issue. https://downloadmirror.intel.com/14687/eng/readme.txt If you have questions, please let us know. Best regards, Michael L. Intel Customer Support Under Contract to Intel Corporation

[Mike_Intel]

Hello MCecc1, I just want to check if you tried our last recommendation and how is the system after trying that? If you have questions, please let us know. Best regards, Michael L. Intel Customer Support Under Contract to Intel Corporation

[Mike_Intel]

Hello MCecc1, I am just sending a follow up to check if you tried our last recommendations and what is the outcome? If you have questions, please let us know. Best regards, Michael L. Intel Customer Support Under Contract to Intel Corporation

[Mike_Intel]

Hello MCecc1, Just want to make another follow up to check if your system is working fine now. If you have questions, please let us know. Best regards, Michael L. Intel Customer Support Under Contract to Intel Corporation

[Mike_Intel]

Hello MCecc1, By the way, you may also contact Cisco Support for further isolation, there might be updated Firmware from Cisco that may resolve this issue. If you have questions, please let us know. Best regards, Michael L. Intel Customer Support Under Contract to Intel Corporation