My company is new to Gitlab.com premium. We almost migrate all of our CI-CD pipeline into gitlab.com(saas version).
Currently i’m trying to migrate my old apps into container.
I try to build my docker images in Gitlab.com Shared Runner to fit this images into AWS ECS.
But after the pull of the "
nvidia/cuda:11.7.1-cudnn8-devel-ubuntu20.04" image, the download of the application binary say that in the dockerfile :
failed to copy files: failed to copy directory: Error processing tar file(exit status 1): write /libtorch_cuda.so: no space left on device
The df -h of the Shared Runner say that :
APP size : 2.2GB
Image Nvidia size : 6.8GB
Is there a solution, or a bypass of this problem on share runner.
or i need to use a private runner for this ?
Thank you for replies,