PowerEdge: XE8640 NVIDIA GDSIO Failed to Run

Summary: Dell PowerEdge XE8640: NVIDIA GDSIO failed to run with an out of memory message.

This article applies to This article does not apply to This article is not tied to any specific product. Not all product versions are identified in this article.

Symptoms

The NVIDIA GPU Direct Storage (GDS) load generator: gdsio fails to run, and the below error messages may be listed:

Kernel: oom_kill_process.cold+0xb/0x10
Kernel: out_of_memory+0xed/0x2e0

Information on GDS and gdsio:
https://docs.nvidia.com/gpudirect-storage/overview-guide/index.htmlThis hyperlink is taking you to a website outside of Dell Technologies.
https://docs.nvidia.com/gpudirect-storage/configuration-guide/index.htmlThis hyperlink is taking you to a website outside of Dell Technologies.

 

Cause

An old version of NVIDIA OpenFabrics Enterprise Distribution (OFED) is installed.

Resolution

Update the OFED to mlnx-ofed-23.07-0.1.0-1 or MOFED greater than 5.9.

Additional Information

This issue was reported on XE8640 only, however, it may also happen on other platforms which support the NVIDIA GDS function.

Affected Products

Mellanox Family of Adapters, PowerEdge XE8640
Article Properties
Article Number: 000221823
Article Type: Solution
Last Modified: 06 May 2025
Version:  3
Find answers to your questions from other Dell users
Support Services
Check if your device is covered by Support Services.