Tag: #unattended-upgrades
Kubernetes, DevOps and MLOps, written from things we ran and measured ourselves · 1 posts
The Day Auto-Update Halted the GPU: New Libraries, Old Kernel Module
On 2026-10-13, a GPU-using pod on a GPU node failed to start with a misleading error about a missing nvidia-persistenced socket. Running nvidia-smi on the node returned Driver/library version mismatch. Three days prior a
2026-10-15 · 9 min read #사후 분석#쿠버네티스#GPU#NVIDIA#Ubuntu