I am speaking of folks on the https://nvidia.custhelp.com site and I have questions for Engineers @ nvidia and others who have undergone the RMA process.
What is the data handling policy for the RMA process? What is best practice? Should I wipe the hard drive / do a system restore before sending in the equipment?
Additionally, my DGX Spark failed the fielddiag for Power Stress with MODS-020000600139. Is there anything I should try before sending it in? Feels like the customer help team is not connected to the knowledge of the product but instead just knows the RMA process.
Others have experienced similar issues. Should I resume the RMA process or am I going to be wasting time by getting another unit with the same issues? Would be great to hear from folks with experience or internal knowledge.
Please open a customer support ticket and they will guide you through the RMA process
So the official reply is that is should do RMA? I’m here worried that I’ll have the same problem or that this might be more of a widespread thing and not fixed by RMA.
If I get a machine through that process, will it be validated against those field tests before it is shipped?
All of our units go through a testing suite after manufacturing. If you get a FAILED log when running Field diagnostics, then you are eligible to RMA
Forgive my skepticism. I know that supply could be constrained and I still don’t know the data handling policy for the RMA process. What will be done with the data on my device? Is there any guarantee of product replacement or is it pending availability? Cause I’d honestly rather not be down a GPU for my personal cluster given how prices are moving.
Once we have received your unit, the drive will be completely wiped and we will send you a new device.