Skiff 2
- vCPU
- 1 × shared
- RAM
- 2 GB
- Storage
- 40 GB NVMe SSD
- Transfer
- 3 TB
- IPv4 / IPv6
- 1 / /64 routed
Operating system
Arch is a rolling release: you always have the newest packages, and you own the consequences. On a server it is a defensible choice if you update deliberately and read the news feed before you do. The AUR gives access to software that is packaged nowhere else.
A KVM virtual server gives Arch Linux its own kernel, its own /proc, and no shared namespace, unlike container-based VPSes. That matters for a rolling distro that ships a stock kernel you can swap for a custom one. You can boot a custom ISO, run nested virtualisation, and use raw sockets. What you sign up for is administration: updates are always current, but you own the consequences. The documentation is outstanding, so the discipline is learnable, but an attentive administrator is not optional.
People run bleeding-edge software that exists nowhere else, often via the AUR. The always-current packages suit personal servers and tools where you want the newest features. The stated weakness is that rolling updates can break things, so workloads here tend to be ones you can rebuild quickly. It is not the choice for a passive, set-and-forget production load.
Before your application takes a byte, Arch needs at least 512 MB just to boot; a 2 GB floor gives you headroom for package updates and the base system's running services. For a minimal server with one or two applications, start at 2 GB. Disk is the cheaper lever: 100 GB NVMe at $8 per month covers package caches and the AUR build directory.
Arch Linux boots in 512 MB and is comfortable with 2 GB. Add whatever your application needs on top of that — the operating system baseline is a floor, not a budget.
Yes, on true KVM virtualisation, which is what every OnionVPS instance runs. It is available as a ready template and can also be installed from your own ISO.
Rolling — no release cycle
The Arch User Repository packages software that is packaged nowhere else, often current development builds. On your own KVM instance you can build from source without restrictions. You trade upstream freshness for occasional build failures, and you verify PKGBUILDs yourself. The documentation covers this thoroughly, which is your safety net.
Yes. True KVM virtualisation gives you your own kernel, so a custom or a hardened kernel works, including one you compiled for your own hardware. No container-based host can do this, because they share the host kernel. You can upload a custom ISO or use the VNC console to manage the boot process.