linux

"Linux is a family of free and open-source software operating systems built around the Linux kernel." - https://en.wikipedia.org/wiki/Linux

Most linux distros are built on GNU tools, and this article is relevant in distinguishing the importance GNU plays in the linux ecosystem: https://www.gnu.org/gnu/why-gnu-linux.en.html

Linux is part of the Unix family tree.

Performance monitoring

Tricks

Best way to see mounts

There are a few ways to see mounts, but most of them will leave out little details in some cases. The best view of mounts is the /proc/self/mountinfo file.

Determine if running kernel is 32 or 64 bit

Works on x86 or ARM.

getconf LONG_BIT

Configure a system to reboot on kernel panic

These lines should be added to sysctl.conf

## Reboot after 10 seconds if kernel panics
kernel.panic = 10
## Treat all oopses as panics
kernel.panic_on_oops = 1

Force reboot on corrupt system

For times that commands like reboot and shutdown are not available.

echo 1 > /proc/sys/kernel/sysrq
echo b > /proc/sysrq-trigger

Show process signals

This should work on other unixes too.

trap -l

Kernel namespaces

"A namespace wraps a global system resource in an abstraction that makes it appear to the processes within the namespace that they have their own isolated instance of the global resource. Changes to the global resource are visible to other processes that are members of the namespace, but are invisible to other processes. One use of namespaces is to implement containers." - man namespaces

"Control cgroups, usually referred to as cgroups, are a Linux kernel feature which allow processes to be organized into hierarchical groups whose usage of various types of resources can then be limited and monitored." - man cgroups

cgroup is one of the linux namespaces. (see man namespaces for more info.)

Tools and stuff

lsns - list namespaces
cgcreate - create new cgroup
cgexec - run the task in given control group
cgclassify - move running task(s) to given cgroup
nsenter - Run a command in a referenced process cgroup config
systemd-cgls - systemd-cgls - Recursively show control group contents
systemd-cgtop - Show top control groups by their resource usage
/proc/self/cgroup - cgroup introspection

Various namespace-aware tool examples

ps cgroup output

ps -o pid,ppid,user,comm,flags,%cpu,sz,%mem,cgname

Run a process in another namespace

With nsenter you specify a target pid to reference, and then specify which namespaces of its you want to enter.

On Ubuntu 18.04, udev mounts devices in a non-global namespace, which prevents normal users from viewing those mounts. You must use nsenter to enter the udevd namespaces to view the mounts, using either --all to get all namespaces of udevd, or --mount for just that one required namespace:

root@bionic:~# lsblk -o NAME,MOUNTPOINT /dev/sdc
NAME   MOUNTPOINT
sdc
└─sdc1

root@bionic:~# nsenter --all -t $(pgrep systemd-udevd) lsblk -o NAME,MOUNTPOINT /dev/sdc
NAME   MOUNTPOINT
sdc
└─sdc1 /mnt/adea64ca-e340-4961-8a4d-75d8a5970664

root@bionic:~# nsenter --mount -t $(pgrep systemd-udevd) lsblk -o NAME,MOUNTPOINT /dev/sdc
NAME   MOUNTPOINT
sdc
└─sdc1 /mnt/adea64ca-e340-4961-8a4d-75d8a5970664

See udev for one permanent fix for this.

Find the path to a namespace

The path to a namespace can be used in some instances instead of the pid. We can discover the path to a namespace by using lsns.

root@bionic:~# lsns -p $(pgrep udevd) -o +PATH
        NS TYPE   NPROCS   PID USER COMMAND                    PATH
4026531835 cgroup    173     1 root /sbin/init                 /proc/1/ns/cgroup
4026531836 pid       173     1 root /sbin/init                 /proc/1/ns/pid
4026531837 user      173     1 root /sbin/init                 /proc/1/ns/user
4026531838 uts       173     1 root /sbin/init                 /proc/1/ns/uts
4026531839 ipc       173     1 root /sbin/init                 /proc/1/ns/ipc
4026532009 net       173     1 root /sbin/init                 /proc/1/ns/net
4026532286 mnt         1  5480 root /lib/systemd/systemd-udevd /proc/5480/ns/mnt