Help Center/ Elastic Cloud Server/ Troubleshooting/ Self-diagnosis of Faulty GPU-accelerated ECSs/ Fault Diagnosis and Handling of Graphics Cards/ How Do I Handle the infoROM Error?
Updated on 2026-01-22 GMT+08:00
View PDF
Share

How Do I Handle the infoROM Error?

Symptom

The error message "WARNING:infoROM is corrupted at gpu 0000:00:0D.0" is displayed when the nvidia-smi command is executed on a Linux ECS, and the services have been affected.

Possible Causes

The health check failed so the GPU driver would not use or trust its content (some content is not used).

Impact

The ECC-related non-volatile data records may be affected. As a result, GPU memory pages that should have been retired are still in use.

Solution

  1. If services are not affected, no operation is required.
  2. If services are affected, stop the services, migrate ECSs, collect fault information by referring to Fault Information Collection, and contact technical support.