Preparation
Changing the kernel on a virtualization host can affect every workload on the node. Before touching the boot configuration, I confirmed that IPMI access was available and kept the existing stable kernel installed as a rollback option.
The node was already running a current stable Proxmox kernel. Liquorix was selected because it is tuned for responsiveness and low-latency desktop and server workloads.
Installation
I installed the latest stable Liquorix packages, refreshed the GRUB configuration and verified that both the new and previous kernels remained available in the boot menu. After restarting the dedicated server, I confirmed the running kernel and checked the Proxmox services, storage and networking before returning workloads to normal operation.
Observation
The useful comparison came after several days rather than immediately after boot. With a similar workload, monitoring showed steadier latency while the node carried more total load. Because kernel behavior is workload-specific, I treated this as an operational result for that node rather than a universal benchmark.
Original references
Original project media
