High Rx checksum errors reported on Intel E810 adapters in NSX environments
Article ID: 423712
Updated On:
Products
Issue/Introduction
ESXi hosts equipped with Intel E810 network adapters may report a high volume of receive (Rx) errors, specifically increasing rapidly (e.g., hundreds of errors every few seconds). These errors are visible in the vSphere Client as pNIC alarms or when monitoring hardware statistics. In environments utilizing VMware NSX, these errors often manifest as checksum failures (csumErr) despite no apparent impact on network performance or packet loss in the upper application layers.
Environment
VMware ESXi 7.x
VMware ESXi 8.x
Cause
The issue is caused by a hardware acceleration interpretation conflict. The adapter incorrectly flags encapsulated packets containing inner packet padding as having incorrect checksums. While the icen driver collects and reports these as hardware errors, the packets are technically valid and continue to be passed to the network stack.
Resolution
Workaround: If an immediate update is not possible, ensure you are monitoring the application layer for actual packet loss. Because these are often "false positives," the performance impact is typically negligible.
Hardware Vendor: Please contact your specific Server OEM (e.g., Dell, HPE, Lenovo) to obtain the validated driver and firmware combo for your specific hardware model
Additional Information
Feedback
