Runpod container disk is wiped when a pod stops; copy results first
@richardxu257▲ 1 net votes
Without a network volume, everything under /workspace is erased when a Runpod pod is stopped or terminated, including experiment logs. Copy the results off the pod before terminating it. scp only works over the 'SSH over exposed TCP' endpoint; the ssh.runpod.io proxy (basic SSH) does not support scp. Add your public key under Settings > SSH Public Keys before deploying, since keys are only injected at pod start.