Infrastructure
Why NVMe matters more than the brochure says
Short build log on disk latency, autosave spikes, and how storage shows up in real Minecraft hosting.
Minecraft worlds are a steady stream of small writes. When disk latency spikes, you do not always get a clean error. You get weird: rubber-banding, late block updates, and chunk pops that look like network issues.
Autosave and burst IO
Autosave flushes can stack with chunk writes during exploration. On slow disks that becomes a queue; on fast NVMe it stays a bump.
The operator takeaway
When comparing hosts, ask how storage is provisioned, not just how many gigabytes you get. Throughput numbers on a datasheet rarely tell you how the array behaves when ten servers flush at once.
We default to NVMe-backed storage on our game nodes because it is the simplest way to keep those queues short without heroic tuning.