server network outage, and the kernel log prompts "kernel i40e xxx: AMD Event logged [IO_PAGE_FAUL

[singless]

Hi all,

I had a problem with two of my servers, there was a network outage, and in the kernel log I saw the error log "i40e 0000:c2:00.1: AMD-Vi: Event logged [IO_PAGE_FAULT". My server is running a ceph cluster. Some information on the Internet shows that the problem may be caused by iommu. My server does have iommu enabled, but my server has been running with iommu enabled for a long time. , so I am not sure whether it is caused by the problem of turning on iommu. In addition, some data show that the network card may encounter this problem when it suffers from a large amount of traffic, but the traffic of my network card was not high at the time. Can anyone help me figure out what might be causing the problem? Thank you so much.

Regards,

OS: Ubuntu 20.04.03 LTS
Kernel: 5.4.0-81-generic
i40e version: 2.8.20-k
X710 firmware: 6.01 0x8000378b 1.1747.0
Server: ESC8000A-E11
CPUs: 2x AMD EPYC 7542 32-Core Processor

links I refer to

https://community.intel.com/t5/Ethernet-Products/i40e-driver-crashes-machine-under-high-network-pressure/m-p/1239793/thread-id/23211

https://askubuntu.com/questions/805008/errors-showing-while-booting-16-04-amd-vi-event-logged-io-page-fault

[singless]

Another host was disconnected from the network at around 14:19 on October 27th.

[Fariz_Intel]

Hi singless,

Thank you for the posting in Intel Ethernet Community.

We understand that the server has been running with IOMMU enabled for a long time and hasn't had any issues. However, for testing purposes, is it possible for you to try disabling it and see if the issue still persists?

Best regards,

Fariz_Intel

[Fariz_Intel]

Hi singless,

If you have additional questions or clarifications, feel free to let us know.

We found some information from the external link below that might be useful in addressing your issue.

https://ubuntuforums.org/showthread.php?t=2290992&page=2&s=36433b0b49c0af0af99af2a317e52d7b

Best regards,

Fariz_Intel

[Fariz_Intel]

Hi singless,

If you require any additional information, please submit a new question, as this thread will no longer be monitored. Other community members may also join in and share their insights on this topic.

Best regards,

Fariz_Intel