Simple note for a Linux newbie
Prevent password request without modifying default configuration file:
echo "$USER ALL=(ALL) NOPASSWD: ALL" | sudo tee /etc/sudoers.d/$USER
On desktop, sshd not installed by default:
sudo apt install openssh-server
sudo systemctl enable ssh
LVM concept in order:
- File systems
- Logical volumes (LVs)
- Volume groups (VGs)
- Physical volumes (PVs) (ie: usually same as partition)
- Partitons
- Disk
Install of Ubuntu server with default option on a 128G disk VM result in only 64G for the /, and missing 64G (lv* need lvm2 package):
$ df -h
Filesystem Size Used Avail Use% Mounted on
tmpfs 784M 700K 783M 1% /run
/dev/mapper/ubuntu--vg-ubuntu--lv 62G 6.6G 52G 12% /
tmpfs 3.9G 0 3.9G 0% /dev/shm
tmpfs 5.0M 0 5.0M 0% /run/lock
/dev/vda2 2.0G 141M 1.7G 8% /boot
/dev/vda1 1.1G 6.4M 1.1G 1% /boot/efi
tmpfs 784M 4.0K 784M 1% /run/user/1000
$ lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINTS
sda 8:0 0 1.9G 1 disk
├─sda1 8:1 0 1.9G 1 part
└─sda2 8:2 0 5.8M 1 part
vda 252:0 0 128G 0 disk
├─vda1 252:1 0 1G 0 part /boot/efi
├─vda2 252:2 0 2G 0 part /boot
└─vda3 252:3 0 124.9G 0 part
└─ubuntu--vg-ubuntu--lv 253:0 0 62.5G 0 lvm /
$ sudo lvdisplay | grep Size
LV Size 62.47 GiB
=> vda3 part is 125G, but lvm uses only 63G Confirmed by the Volume group "Free space":
$ sudo vgdisplay | grep Free
Free PE / Size 15993 / 62.47 GiB
So need to extend it:
sudo lvextend -l +100%FREE /dev/ubuntu-vg/ubuntu-lv
$ sudo lvdisplay | grep Size
LV Size <124.95 GiB
And the filesystem too:
sudo resize2fs /dev/mapper/ubuntu--vg-ubuntu--lv
$ df -h | grep ubuntu
/dev/mapper/ubuntu--vg-ubuntu--lv 123G 6.6G 111G 6% /
Warning: Ubuntu uses netplan and this mess uses:
- NetworkManager on Ubuntu Desktop
- (systemd) networkd on Ubuntu Server So stick to netplan (/etc/netplan/01-netcfg.yaml) on those. If this file is missing, create it:
netplan gen > /etc/netplan/01-netcfg.yaml
chmod 600 /etc/netplan/01-netcfg.yaml
And edit the renderer line to the installed API (switching back to networkd if you’ve upgraded a server to desktop without installing NetworkManager as example).
A mess between the Network-manager (nmcli) and systemd-resolve.
Interface status:
nmcli device status
Displaying DHCP info:
nmcli -f DHCP4
nmcli device show enp2s0
What is the DHCP server IP address ?
nmcli -f DHCP4 connection show "Wired connection 1"
What is the DNS server ?
resolvectl status
Static IP configuration in /etc/netplan/02-netconfig.yaml
Official docs:
A debian packages is build with 3 files:
- name_version.dsc (description text file)
- name_version.orig.tar.xz (original sources archives)
- name_version.debian.tar.xz (debian patches and build scripts, to be untar into the sources dir)
Generic linux tooling needed:
- build-essential (compilers)
- devscript (packages build tools, like dpkg-depcheck)
- packages listed in Build-Depends field
- packages listed in the Build-Depends-indep field of debian/control
About the package build dependencies, file debian/control should be the one, but the dpkg-buildpackage will display all missings deps.
Do not (download original sources)[https://www.kernel.org/pub/linux/utils/util-linux/], but the Ubuntu repository fork.
git clone -b ubuntu/jammy --single-branch https://git.launchpad.net/ubuntu/+source/util-linux
cd util-linux
=> Install build-deps from debian/control
DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage -b
==== 2. Install build dependencies
Reading the debian/control file gives more clue: sudo apt install bison libaudit-dev libcap-ng-dev libcrypt-dev libcryptsetup-dev libncurses5-dev libncursesw5-dev libpam0g-dev libreadline-dev libselinux1-dev libsystemd-dev libtool libudev-dev zlib1g-dev libaudit-dev
Then running (it will display the exact list of missing build dependencies too):
DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage -b
DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage -us -uc -b
-us, --unsigned-source unsigned source package.
-uc, --unsigned-changes unsigned .buildinfo and .changes file.
-b, --build=binary binary-only, no source files.
env nocheck: avoid running regression test
Ubuntu is no more 'clean' and adding crap like ESM and snap.
So, a lot of work need to be done:
- Replace ubuntu-advantage-tools with a fake package
- Disable ESM motd spam
- Disable dynamic motd spam
sudo pro config set apt_news=false
wget -q --content-disposition https://github.com/Skyedra/UnspamifyUbuntu/blob/master/fake-ubuntu-advantage-tools/fake-ubuntu-advantage-tools.deb?raw=true
sudo apt install ./fake-ubuntu-advantage-tools.deb
Hidden repo (not showed with add-apt-repository --list), that need to be manually disabled
sed -i 's/^deb/#deb/g' /var/lib/ubuntu-advantage/apt-esm/etc/apt/sources.list.d/ubuntu-esm-apps.list
snap list
snap remove --purge packages-in-the-list
apt remove snapd
apt purge snapd
Disable privacy address:
- Replace net.ipv6.conf.*.use_tempaddr = 2 in /etc/sysctl.d/10-ipv6-privacy.conf buy = 0
Display DNS parameters:
resolvectl
Flush DNS cache:
sudo resolvectl flush-caches
lsblk
sudo lshw
lshw -c video
lspci
listing devices that need a drivers:
ubuntu-drivers devices
If error:
$ ubuntu-drivers devices
ERROR:root:aplay command not found
Then install alsa-utils:
sudo apt-get install -y alsa-utils
sudo apt install -y mesa-utils
DISPLAY=:0 glxinfo -B
If error:
Error: unable to open display :0
$ sudo apt-get install -y libze-intel-gpu1 libze1 intel-ocloc intel-opencl-icd clinfo
$ clinfo -l
Platform #0: Intel(R) OpenCL Graphics
`-- Device #0: Intel(R) Graphics [0x7d55]
$ lshw -c video
74:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Rembrandt (rev 0a)
$ sudo dmesg | egrep 'drm|radeon'
(etc.)
[ 2.317230] [drm] Initialized amdgpu 3.48.0 20150101 for 0000:74:00.0 on minor 0
(etc)
Binary from AMD website and install doc. Check on the repository the latest version
curl -O https://repo.radeon.com/amdgpu-install/23.10.2/ubuntu/jammy/amdgpu-install_5.5.50502-1_all.deb
sudo apt-get install ./amdgpu-install_5.4.50502-1_all.deb
sudo amdgpu-install --usecase=graphics --vulkan=amdvlk --opencl=rocr
sudo usermod -a -G render $LOGNAME
sudo usermod -a -G video $LOGNAME
echo "VK_ICD_FILENAMES=/etc/alternatives/amd_icd64.json" | sudo tee -a /etc/environment
sudo reboot
If proprietary drivers needed, replace amdgpu by this line:
sudo amdgpu-install --usecase=graphics --vulkan=pro --opencl=rocr
Need to test Vulkan API with vkcube tool:
sudo apt-get install vulkan-tools
vulkaninfo
vkcube
Use radeontop to see GPU usage
If already installed by ubuntu-drivers autoinstall, need to be removed first:
apt remove nvidia-headless-525-server nvidia-dkms-525-server nvidia-headless-no-dkms-525-server nvidia-kernel-common-525
wget https://us.download.nvidia.com/XFree86/aarch64/530.41.03/NVIDIA-Linux-aarch64-530.41.03.run
sh ./NVIDIA-Linux-aarch64-530.41.03.run
Warning: CUDA version MUST match minimum drivers version following CUDA Install guide Testing:
$ nvidia-smi
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 530.41.03 Driver Version: 530.41.03 CUDA Version: 12.1 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 Tesla T4 Off| 00000000:01:00.0 Off | 0 |
| N/A 58C P8 10W / 70W| 2MiB / 15360MiB | 0% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
| 1 Tesla T4 Off| 00000001:01:00.0 Off | 0 |
| N/A 63C P8 11W / 70W| 2MiB / 15360MiB | 0% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
| 2 Tesla T4 Off| 00000007:01:00.0 Off | 0 |
| N/A 56C P8 11W / 70W| 2MiB / 15360MiB | 0% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| No running processes found |
+---------------------------------------------------------------------------------------+
Running hascat in bench mode (-b):
$ nvidia-smi
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 530.41.03 Driver Version: 530.41.03 CUDA Version: 12.1 |
|-----------------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+======================+======================|
| 0 Tesla T4 Off| 00000000:01:00.0 Off | 0 |
| N/A 70C P0 72W / 70W| 2984MiB / 15360MiB | 99% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
| 1 Tesla T4 Off| 00000001:01:00.0 Off | 0 |
| N/A 71C P0 71W / 70W| 2984MiB / 15360MiB | 99% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
| 2 Tesla T4 Off| 00000007:01:00.0 Off | 0 |
| N/A 65C P0 75W / 70W| 2984MiB / 15360MiB | 100% Default |
| | | N/A |
+-----------------------------------------+----------------------+----------------------+
+---------------------------------------------------------------------------------------+
|7 Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=======================================================================================|
| 0 N/A N/A 12914 C hashcat 2980MiB |
| 1 N/A N/A 12914 C hashcat 2980MiB |
| 2 N/A N/A 12914 C hashcat 2980MiB |
+---------------------------------------------------------------------------------------+
Last action: shrink the Windows NTFS volume, and expanded the Linux ext4fs rootfs volume. System was working great, but after the first reboot it refuse to boot.
(initramfs) blkid
/dev/nvme0n1p3: UUID=xxxx TYPE="ntfs" PARTLABEL="Basic data partiton"
/dev/nvme0n1p1: UUID=yyyy LABEL="SYSTEM" TYPE="vfat" PARTLABEL="EFI"
/dev/nvme0n1p4: UUID=zzzz LABEL="Recovery" TYPE="nts"
/dev/nvme0n1p2: UUID=wwww, PARTLABEL="Microsotf reserver partition"
(same for an ls /dev/disk/by-uuid/)
Here, it is a dual boot Windows/Ubuntu, but where is my extfs partition ??
(initramfs) ls /dev/nvme0n"*
/dev/nvme0n1
/dev/nvme0n1p1
/dev/nvme0n1p2
/dev/nvme0n1p3
/dev/nvme0n1p4
/dev/nvme0n1p5
Partition 5 is missing from the blkid output, let’s force an fsck:
fsck.ext4 /dev/nvme0n1p5
fsck doesn’t report any errror, but why no UUID displayed by blkid ? Let’s try to run tunefs, but it is on the broken partition:
(initramfs) mkdir /mnt
(initramfs) mount /dev/nvme0n1p5 /mnt
(initramfs) /mnt/sbin/tune2fs -l /dev/nvme0n1p5 | grep UUID
Filesystem UUID: 1db6b810-d625-4e9a-aced-32b48f6a8d5
What ?? It is here! Ok, let’s replace the UUID usage by the disk name in /etc/fstab:
(initramfs) mount --bind /dev /mnt/dev
(initramfs) mount --bind /proc /mnt/proc
(initramfs) chroot /mnt
root@(none):/# vi /etc/fstab
Here I’ve replaced the entry /dev/disk/by-uuid/1dd... by /dev/nvme0n1p5
Now upgrading all bootloader and initramfs then reboot:
root@(none):/# update-grub
root@(none):/# update-initramfs -u
root@(none):/# exit
(initramfs) umount /mnt/proc
(initramfs) umount /mnt/dev
(initramfs) umount /mnt
(initramfs) reboot -f
And still not able to find the now /dev/nvme0n1p5 during boot. Let’s try to pass a longer rootdelay in Linux kernel by updating grub. Remount and re-chroot into your rootfs, then set a 10 seconds delay to wait to detect this disk:
vi /etc/default/grub
GRUB_CMDLINE_LINUX="rootdelay=10"
Then redo the same update-grub+exit+umount+reboot steps as previously. But... still nothing. So booting with an USB live CD:
fdisk -l /dev/nvme0n1
All partitons display, including p5 as "Linux filesystem" type
blkid /dev/nvme0n1p5
Empty reply, let’s probe it:
sudo blkid --probe --match-types ext4 /dev/nvme0n1p5
/dev/nvme0n1p5: UUID="1dd..."
Let’s check if partprobe error:
sudo partprobe /dev/nvme0n1p5
No error displayed but still nothing, so let’s change the UUID:
sudo tune2fs -U random /dev/nvme0n1p5
Setting the UUID on this filesystem could take some time.
Proceed anyway (or wait 5 seconds to proceed) ? (y,N)
Still not fix, so let’s change the disk partition number (p4 position if after p5, so should change p5 as p4 and p4 as p5):
sudo fdisk /dev/nvme0n1
p (print partition table)
x (extra functionnality)
f (fix partition order)
r (return to main menu)
p (print partition table to check new order)
w (write and exit)
Now time to mount disk and chroot into it to update /etc/fstab with new partition id. While here I’ve noticed this error message "couldn’t identify filesystem type for fsck hook, ignoring" from update-initramfs. It was not able to automatically detected this partition. Solution: reformating.
It creates a swap file in the /:
ubuntu:/# swapon -show
Filename Type Size Used Priority
/swap.img file 8388604 0 -2
To disable it:
swapoff /swap.img
rm /swap.img
sed -i '/swap.img/d' /etc/fstab
reverse mode:
sudo journalctl -r
Options: -i install -r remove -P purge (remove config file) -l list --force-all
From which package this file belong to ?
dpkg -S /usr/bin/m4
If need multiples compilers:
update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-10 10
update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-10 10
update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-11 11
update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-11 11
Now to select them:
update-alternatives --config g++
update-alternatives --config gcc
Switch the 1 to 0 here:
sudo vi /etc/apt/apt.conf.d/20auto-upgrades
APT::Periodic::Update-Package-Lists "1";
APT::Periodic::Unattended-Upgrade "1";
apt-get install linux-tools-common linux-tools-generic linux-tools-`uname -r`
perf record -f -g -a
sudo usermod -aG docker $USER
newgrp docker
docker run hello-world
Where are image stored:
$ docker info
...
Storage Driver: overlay2
Docker Root Dir: /var/lib/docker
Customize a docker image:
olivier@host:~$ docker run -it ubuntu bash
root@f18e5dff2f03:/# lsb_release -a
bash: lsb_release: command not found
root@f18e5dff2f03:/# apt update && apt install lsb-core
(etc.)
root@f18e5dff2f03:/# lsb_release -a
LSB Version: core-11.1.0ubuntu4-noarch:security-11.1.0ubuntu4-noarch
Distributor ID: Ubuntu
Description: Ubuntu 22.04.2 LTS
Release: 22.04
=> on other terminal
olivier@host~$ docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
f18e5dff2f03 ubuntu "bash" 6 minutes ago Up 6 minutes goofy_almeida
olivier@host:~$ docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
ubuntuwithlsb latest bae41ca408fc 7 seconds ago 586MB
hello-world latest 9c7a54a9a43c 2 weeks ago 13.3kB
ubuntu latest 3b418d7b466a 3 weeks ago 77.8MB
olivier@host:~$ docker run -it ubuntuwithlsb bash
root@d781574512cd:/# lsb_release -a
LSB Version: core-11.1.0ubuntu4-noarch:security-11.1.0ubuntu4-noarch
Distributor ID: Ubuntu
Description: Ubuntu 22.04.2 LTS
Release: 22.04
Codename: jammy
Persisting data folder:
sudo docker run -ti --rm -v ~/Docker_Share:/data ubuntu /bin/bash
docker container ls
docker ps --no-trunc
Copying files:
docker cp ./some_file CONTAINER:/work
Open shell inside:
docker exec -it CONTAINER sh
System stat:
docker stats
Cleaning docker layer cache (can be in weird buggy state):
yes | docker system prune -a
sudo apt install linux-tools-generic
Example:
olivier@ryzen7:~$ sudo perf record --call-graph dwarf /usr/bin/radeontop -d - -i 15 -t 20
Dumping to -, until termination.
1683580703.217021: bus 74, gpu 0.00%, ee 0.00%, vgt 0.00%, ta 0.00%, sx 0.00%, sh 0.00%, spi 0.00%, sc 0.00%, pa 0.00%, db 0.00%, cb 0.00%, vram 2.97% 242.41mb, gtt 0.06% 17.50mb, mclk 43.61% 1.047ghz, sclk 18.18% 0.400ghz
1683580718.217399: bus 74, gpu 0.00%, ee 0.00%, vgt 0.00%, ta 0.00%, sx 0.00%, sh 0.00%, spi 0.00%, sc 0.00%, pa 0.00%, db 0.00%, cb 0.00%, vram 2.97% 242.41mb, gtt 0.06% 17.50mb, mclk 43.61% 1.047ghz, sclk 18.18% 0.400ghz
1683580733.217864: bus 74, gpu 0.00%, ee 0.00%, vgt 0.00%, ta 0.00%, sx 0.00%, sh 0.00%, spi 0.00%, sc 0.00%, pa 0.00%, db 0.00%, cb 0.00%, vram 2.97% 242.41mb, gtt 0.06% 17.50mb, mclk 42.06% 1.009ghz, sclk 18.18% 0.400ghz
1683580748.218396: bus 74, gpu 0.00%, ee 0.00%, vgt 0.00%, ta 0.00%, sx 0.00%, sh 0.00%, spi 0.00%, sc 0.00%, pa 0.00%, db 0.00%, cb 0.00%, vram 2.97% 242.41mb, gtt 0.06% 17.50mb, mclk 43.42% 1.042ghz, sclk 18.18% 0.400ghz
1683580763.218933: bus 74, gpu 0.00%, ee 0.00%, vgt 0.00%, ta 0.00%, sx 0.00%, sh 0.00%, spi 0.00%, sc 0.00%, pa 0.00%, db 0.00%, cb 0.00%, vram 2.97% 242.41mb, gtt 0.06% 17.50mb, mclk 42.64% 1.023ghz, sclk 18.18% 0.400ghz
^C1683580765.868285: bus 74, gpu 0.00%, ee 0.00%, vgt 0.00%, ta 0.00%, sx 0.00%, sh 0.00%, spi 0.00%, sc 0.00%, pa 0.00%, db 0.00%, cb 0.00%, vram 2.97% 242.41mb, gtt 0.06% 17.50mb, mclk 42.44% 1.019ghz, sclk 18.18% 0.400ghz
[ perf record: Woken up 36 times to write data ]
[ perf record: Captured and wrote 9.256 MB perf.data (1136 samples) ]
systemctl get-default
If it displays multi-user.target (text) switch it to graphical:
sudo systemctl set-default graphical.target
sudo sed -i 's/#WaylandEnable/WaylandEnable/' /etc/gdm3/custom.conf
sudo systemctl restart gdm3
To be able to start graphical software from SSH, need to export XAUTHORITY and DISPLAY. Your user need to be locally logged on the wayland/xorg (to create the XAUTH token file), then from SSH session:
export XAUTHORITY=$(ls /run/user/$(id -u)/.* | grep auth)
export DISPLAY=:0
Then try to run simple app:
xcalc