Virtctl upload memory requirments

[mw-0]

What happened:
When using the import feature " virtctl image-upload dv ......" i get an error when it hits the ram limit. When i try and upload a larger 5GB qcow image i have to set the podresourcerequirements in the cdi spec to a higher amount or memory than the download size.

I have tested a few different images and found the same issue when matching the limit lower.

Environment:

• CDI version (use kubectl get deployments cdi-deployment -o yaml): 1.61.0
• Kubernetes version (use kubectl version): 1.31.2
• OS (e.g. from /etc/os-release): Debian12
• Kernel (e.g. uname -a): 6.10.11

[akalenyu]

So you are seeing OOMs on the upload pod? which storage class are you using?
We've had similar situations with some kernel bug, was very common with nfs based storage classes:
https://gitlab.com/redhat/centos-stream/src/kernel/centos-stream-9/-/merge_requests/4523

[mw-0]

i am using NFS 4.1 PVC with Netapp Trident. I can see the pvc are attached. Is the patch upstream into kernel or specific for RHEL there?

[akalenyu]

It made it into centos stream. But I don't know where it stands in other distros. If you have the debian sources should be easy enough to check if the revert made it in.