Skip to main content

Dedicated Servers

40 articles in this section

A dedicated server differs from a VDS in one key way: the hardware is entirely yours — CPU, disks, and network are not shared with anyone else. This section starts with the choices you make when ordering: how to judge a processor — Xeon, EPYC, or Ryzen — why ECC memory matters, how NVMe compares to SATA SSD and HDD, which RAID level fits a given workload, and when a hardware controller beats mdadm.

From there it moves to running hardware you already lease: SMART diagnostics for drives, replacing a disk in a RAID array without downtime, benchmarking with fio, sysbench, and iperf3, remote management through IPMI and KVM-over-IP, installing an OS via Virtual Media, and recovering through rescue mode when a server won't boot. A separate guide covers acceptance checks for the first hour after you get access.

There are also decision-making guides: colocation versus renting, estimating the cost of a dedicated server, choosing a datacenter location, the 3-2-1 backup scheme, and hardware requirements for databases, gaming projects, video streaming, and GPU-based model training.

  • Choosing a CPU, memory, and disks for the workload
  • RAID, ZFS, and the disk subsystem
  • IPMI, rescue mode, and hardware monitoring
  • Datacenter SLAs and the 3-2-1 backup scheme
What a Dedicated Server Is and When You Need One A dedicated server is a whole physical machine you rent: hardware breakdown, real use cases, and the signals that tell you a VPS no longer carries your load. Dedicated Server or VPS: Choosing by Real Workload Dedicated server or VPS: compared by steal time, disk latency and cost, with the commands to measure your load and the point where bare metal gets cheaper. How to Choose a Server CPU: Xeon, EPYC and Ryzen How to choose a server CPU: when clock speed matters, when core count wins, how Xeon, EPYC and Ryzen differ, and how to verify real delivered performance. ECC Memory in a Server: Why It Matters and How to Check ECC memory corrects single-bit errors and detects double-bit ones: the difference from plain RAM, the commands that verify it, and how to read CE counters. NVMe, SATA SSD or HDD: Picking Drives for a Server NVMe, SATA SSD or HDD in a server: compared by IOPS, latency and write endurance, matched to databases, file storage and archives, with fio test commands. RAID 0, 1, 5, 10: Which Level to Choose for a Server RAID 0, 1, 5 and 10 on a dedicated server: fault tolerance, usable capacity, write penalty and rebuild time, plus how to pick a level for your workload. Hardware RAID or mdadm: What to Choose on a Server Hardware RAID versus mdadm on a dedicated server: write cache, CPU load, disk portability, monitoring tools and reading SMART behind a controller. SMART Disk Diagnostics on a Server: smartctl and NVMe SMART disk diagnostics on a server with smartctl: which attributes decide, how to read the NVMe health log, run self-tests and set up smartd for alerts. Replacing a Disk in a RAID Array with No Downtime Replacing a disk in a RAID array step by step: removing it with mdadm, cloning the layout, adding it back, watching the rebuild and installing the bootloader. Benchmarking a Dedicated Server: fio, sysbench, iperf3 Benchmarking a dedicated server: fio profiles for disks, sysbench for CPU and memory, iperf3 for the uplink, plus how to read metrics and avoid test mistakes. IPMI and KVM-over-IP: out-of-band server management IPMI and KVM-over-IP on a dedicated server: what the BMC does, how to connect, ipmitool commands, Serial-over-LAN, and why the management port stays private. Installing an OS on a Dedicated Server via IPMI Install an OS on a dedicated server over IPMI: mount the ISO, set boot order with ipmitool, partition for RAID, automate with preseed or kickstart, verify. Rescue mode: recovering a server that will not boot Rescue mode on a dedicated server: assemble RAID and LVM, mount the root, enter a chroot, repair fstab, GRUB and initramfs, and bring the machine back online. Server hardware monitoring: temperature, disks, power Hardware monitoring for a dedicated server: IPMI sensors and the SEL, SMART, EDAC counters, RAID state, exporting metrics to Prometheus and alert thresholds. Dedicated server acceptance: what to check in the first hour Dedicated server acceptance: verify CPU, ECC memory, disks and uplink against the order, check SMART and the array, and lock access before migrating. Server RAM testing: memtest86+ and EDAC counters How to test server RAM: EDAC counters in sysfs on a live machine, memtester without a reboot, and a full memtest86+ run mounted over IPMI from an ISO. Server BIOS/UEFI setup: performance and boot options Dedicated server BIOS/UEFI setup: boot mode, power profile, C-states, VT-x and VT-d, boot order with efibootmgr and how to verify it all from Linux. Server uplink: Mbps, unmetered plans and traffic caps How unmetered differs from a terabyte cap, how to convert Mbps into monthly TB, how to measure the uplink with iperf3 and track traffic using vnstat. Additional IPv4 addresses and subnets on a server Extra IPv4 on a dedicated server: $5/month pricing, single address versus a /29 subnet, netplan and nmcli setup, PTR records and outbound source IP. DDoS protection for a dedicated server: L3, L4, L7 DDoS protection for a dedicated server: how L3, L4 and L7 differ, what is included in Germany, the USA and France, plus sysctl, nftables and nginx limits. Dedicated Server Location: Germany, USA or France How to choose a dedicated server location between Germany, the USA and France: ping and traceroute methodology, uplink, DDoS protection and plan prices. Virtualization on a Server: KVM, Proxmox or ESXi Virtualization on a dedicated server: KVM, Proxmox VE and ESXi compared, VT-x checks, bridge setup, guest density per plan and the usual mistakes. Dedicated Server for a Database: Hardware and Tuning A dedicated server for a database: how much RAM and which disks, InnoDB and PostgreSQL tuning, fio latency testing and choosing the right RAID level. Game Server Hardware: Clock Speed Beats Core Count Game server hardware: why core clock decides tick time, how much RAM per instance, why NVMe for saves, and how to tune Linux for a steady tick rate. Dedicated Server for Video and Streaming: Hardware Guide A dedicated server for video and streaming: sizing the uplink by bitrate and viewers, measuring transcoding with ffmpeg, HLS packaging and archive storage. Server for 1C: hardware and configuration requirements What server 1C needs: core frequency over core count, ECC memory sizing, NVMe for the database, Windows Server with MS SQL or Linux with PostgreSQL. What makes up the cost of a dedicated server rental The cost of a dedicated server broken down: hardware, uplink, IP and site in plans from $49 to $149, plus IPv4, KVM/IPMI and backup options. Colocation or renting a server: which one pays off Colocation versus renting a server: compared by upfront cost, hardware failures, upgrades and break-even, with the maths over a three-year operating horizon. Migrating a project from VPS to a dedicated server Migrating a project from VPS to a dedicated server step by step: inventory, rsync and database dump, curl checks by IP, and a DNS switch with a short TTL. Backing up a dedicated server: the 3-2-1 scheme Backing up a dedicated server: what to copy and how often, restic and borg, database dumps, external storage and a monthly restore test that actually runs. Server Form Factor 1U vs 2U: What to Pick Server form factor 1U or 2U affects cooling, drive count, and rack unit price — the differences and how to choose explained. Bonding and LACP: Combining Server Network Ports Bonding and LACP on a dedicated server boost channel speed and protect against a port failure — modes, setup, and verification. ZFS on a Dedicated Server: Pools, ARC, Snapshots ZFS on a dedicated server combines RAID, ARC caching, and snapshots in one layer — pool structure, memory use, and backups. Disk Subsystem: Queue Depth, IOPS, and Latency A server disk subsystem under load hits its I/O queue limit — IOPS, latency, iostat, and fio testing explained step by step. Server Power: Dual PSU, PDU, and Redundancy Server power redundancy protects against a single PSU failure — the N+1 scheme, separate PDUs, power sizing, and testing. Data Center SLA and Tier: What 99.9% and 99.99% Mean Data center SLA and Tier levels reveal real server reliability: Tier I-IV, downtime in minutes, and what the provider guarantee excludes. Updating Server Firmware: BIOS, BMC, and Drives Updating server firmware for BIOS, BMC, and drives closes vulnerabilities and sensor errors: update order, risks, and commands to check versions first. NUMA on Dual-CPU Servers: Process and Memory Pinning NUMA on a dual-processor server affects memory latency: we cover node topology, binding processes with numactl, and common database configuration mistakes. Dedicated GPU Server: How to Choose a Card for ML A dedicated GPU server requires calculating VRAM, power, and PCIe lanes: choosing a card for training, inference, rendering, and passthrough virtualization. Server Event Log (SEL): Reading and Diagnosing Failures The server SEL event log stores failures in the BMC independent of the OS: we cover ipmitool commands, typical ECC and power entries, and when to clear the log.