Check CPU, RAM, disks, and network

After Rescue or a fresh install, confirm you received the hardware you paid for. These commands work the same on Ubuntu, Debian, and the Hetzner Rescue System.

One-shot overview

Identity and resources

root@server ~ # hostnamectlroot@server ~ # lscpuroot@server ~ # free -hroot@server ~ # lsblkroot@server ~ # ip -br a

What you should see

 Static hostname: dedicated       Operating System: Ubuntu 24.04.1 LTS                 Kernel: Linux 6.8.0-41-generic           Architecture: x86-64 Architecture:                            x86_64CPU(s):                                  32Model name:                              AMD EPYC 7502P 32-Core ProcessorVirtualization:                          AMD-V                total        used        free      shared  buff/cache   availableMem:           125Gi        4.1Gi        98Gi        12Mi        23Gi        120GiSwap:          8.0Gi          0B       8.0Gi NAME        SIZE TYPE MOUNTPOINTnvme0n1   476.9G disknvme1n1   476.9G disk lo               UNKNOWN  127.0.0.1/8enp35s0          UP       203.0.113.10/26  2a01:4f8:xxx::2/64

CPU

Cores, model, virtualization flags

root@server ~ # lscpuroot@server ~ # nprocroot@server ~ # lscpu -eroot@server ~ # grep -E 'model name|cpu cores|siblings|Flags' /proc/cpuinfo | headroot@server ~ # egrep -c '(vmx|svm)' /proc/cpuinfo

What you should see

32CPU NODE SOCKET CORE L1d:L1i:L2:L3 ONLINE  0    0      0    0 0:0:0:0          yes  1    0      0    1 1:1:1:0          yes...model name	: AMD EPYC 7502P 32-Core Processor32

vmx (Intel) or svm (AMD) means the CPU can run virtual machines — required later for Proxmox / KVM. A count of 0 means virtualization is off in BIOS (ask Hetzner / check Robot if you expected it).

Note

lscpu “CPU(s)” is threads. On SMT/Hyper-Threading, that is often 2× the physical cores.

RAM

Installed memory

root@server ~ # free -hroot@server ~ # cat /proc/meminfo | headroot@server ~ # dmidecode -t memory | grep -E 'Size:|Speed:|Manufacturer:'

What you should see

               total        used        free      shared  buff/cache   availableMem:           125Gi        4.1Gi        98Gi        12Mi        23Gi        120GiSwap:          8.0Gi          0B       8.0Gi MemTotal:       131686400 kBMemFree:        103809024 kBMemAvailable:   126877696 kB 	Size: 32 GB	Speed: 3200 MT/s	Manufacturer: Samsung

dmidecode reads firmware tables. Install it if missing: apt install dmidecode. In Rescue it is often already there.

Disks

Hardware stack from NVMe disk through partitions, RAID, and LVM — Picture: when you run lsblk you are reading this stack. Confirm the size and that ROTA is 0 for NVMe/SSD.

Capacity, type, rotation vs NVMe

root@server ~ # lsblk -d -o NAME,SIZE,MODEL,ROTA,TRANroot@server ~ # lsblk -o NAME,SIZE,TYPE,FSTYPE,MOUNTPOINTroot@server ~ # fdisk -l# NVMe list (if nvme-cli is installed)root@server ~ # nvme list

What you should see

NAME      SIZE MODEL                    ROTA TRANnvme0n1 476.9G SAMSUNG MZVL2512         0    nvmenvme1n1 476.9G SAMSUNG MZVL2512         0    nvme Node             SN                   Model                    Namespace  Usage/dev/nvme0n1     S4XXXXXX             SAMSUNG MZVL2512         1          512 GB/dev/nvme1n1     S4YYYYYY             SAMSUNG MZVL2512         1          512 GB

ROTA 0 is SSD/NVMe. ROTA 1 is a spinning HDD. TRAN shows sata, nvme, sas.

Disk health (SMART)

root@server ~ # apt install -y smartmontoolsroot@server ~ # smartctl -a /dev/sdaroot@server ~ # smartctl -a /dev/nvme0n1

What you should see

SMART overall-health self-assessment test result: PASSEDTemperature:                        34 CelsiusAvailable Spare:                    100%Percentage Used:                    3%Power On Hours:                     4120

Tip

Look at overall health, reallocated sectors, and temperature. A failing SMART status is a reason to snapshot data and open a Hetzner support ticket for a disk swap.

PCI devices and RAID cards

What is plugged in

root@server ~ # lspciroot@server ~ # lspci | grep -i -E 'raid|nvme|ethernet|vga'root@server ~ # lshw -short

What you should see

01:00.0 Non-Volatile memory controller: Samsung Electronics Co Ltd NVMe SSD Controller01:00.1 Non-Volatile memory controller: Samsung Electronics Co Ltd NVMe SSD Controller35:00.0 Ethernet controller: Intel Corporation Ethernet Controller 10G X550T

If you see Adaptec, LSI, or MegaRAID, that is hardware RAID. Configure it before installimage. If you only see NVMe or AHCI SATA, use software RAID.

Network

Hetzner gives you a public IPv4 (and usually IPv6). The interface name may be eth0, enp*, or similar.

Addresses and routes

root@server ~ # ip -br aroot@server ~ # ip -4 aroot@server ~ # ip -6 aroot@server ~ # ip rroot@server ~ # ip -6 rroot@server ~ # cat /etc/network/interfacesroot@server ~ # ls /etc/netplan 2>/dev/null

What you should see

lo               UNKNOWN  127.0.0.1/8enp35s0          UP       203.0.113.10/26 default via 203.0.113.1 dev enp35s0203.0.113.0/26 dev enp35s0 proto kernel scope link src 203.0.113.10 01-netcfg.yaml

Debian often uses /etc/network/interfaces. Ubuntu often uses Netplan under /etc/netplan. Rescue can show live addresses even when those files are empty.

Temperatures and load

Quick health

root@server ~ # uptimeroot@server ~ # cat /proc/loadavgroot@server ~ # apt install -y lm-sensorsroot@server ~ # sensors

What you should see

 10:22:01 up 14 days,  3:11,  1 user,  load average: 0.08, 0.12, 0.090.08 0.12 0.09 1/428 18420 k10temp-pci-00c3Adapter: PCI adapterTctl:         +41.1°C