Linux

These are the most common commands I use if my systemd service file is placed in /etc/systemd/system/myservice.service.

Enable (i.e. start at boot) and also start the service right now (--now):

sudo systemctl enable --now myservice

Start by

sudo systemctl start myservice

Restart by

sudo systemctl restart myservice

Stop by

sudo systemctl stop myservice

View status:

sudo systemctl status myservice

View & follow logs:

sudo journalctl -xfu myservice

View logs in less:

sudo journalctl -xu myservice

Problem:

My .jitsi-meet-cfg/jigasi/sip-communicator.properties is getting overwritten every time I start Jigasi, but I need to set

net.java.sip.communicator.impl.protocol.sip.acc1.AUTHORIZATION_NAME=abc123abc

in order for my SIP communication to work.

Solution:

Run this script after starting the jigasi container. It will fix the overwritten config and then restart the Jigasi Java process without restarting the container

#!/bin/sh

sed -i -e "s/# SIP account/net.java.sip.communicator.impl.protocol.sip.acc1.AUTHORIZATION_NAME=abc123abc/g" .jitsi-meet-cfg/jigasi/sip-communicator.properties

# Reload config hack
docker-compose -f docker-compose.yml -f jigasi.yml exec jigasi /bin/bash -c 'kill $(pidof java)'

Original source: This GitHub ticket which provides a similar solution for a similar problem

A good start is to use -P 4 -n 1 to run 4 processes in parallel (-P 4), but give each instance of the command to be run just one argument (-n 1)

These are the xargs options for parallel use from the xargs manpage:

-P, --max-procs=MAX-PROCS    run at most MAX-PROCS processes at a time

-n, --max-args=MAX-ARGS      use at most MAX-ARGS arguments per command line

Example:

cat urls.txt | xargs -P 4 -n 1 wget

This command will run up to 4 wget processes in parallel until each of the URLs in urls.txt has been downloaded. These processes would be run in parallel

wget [URL #1]
wget [URL #2]
wget [URL #3]
wget [URL #4]

If you would use -P 4 -n 2 these processes would be run in parallel:

wget [URL #1] [URL #2]
wget [URL #3] [URL #4]
wget [URL #5] [URL #6]
wget [URL #7] [URL #8]

Using a higher value for -n might slightly increase the efficiency since fewer processes need to be initialized, but it might not work with some commands if you pass multiple arguments.

You can use a regular expression to grep for href="..." attributes in a HTML like this:

grep -oP "(HREF|href)=\"\K.+?(?=\")"

grep is operated with -o (only print match, this is required to get extra features like lookahead assertions) and -P (use Perl regular expression engine). The regular expression is basically

href=".*"

where the .+ is used in non-greedy mode (.+?):

href=".+?"

This will give us hits like

href="/files/image.png"

Since we only want the content of the quotes (") and not  the href="..." part, we can use positive lookbehind assertions (\K) to remove the href part:

href=\"\K.+?\"

but we also want to get rid of the closing double quote. In order to do this, we can use positive lookahead assertions ((?=\")):

href=\"\K.+?(?=\")

Now we want to match both href and HREF to get some case insensitivity:

(href|HREF)=\"\K.+?(?=\")

Often we want to specifically match one file type. For example, we could match only .png:

(href|HREF)=\"\K.+?\.png(?=\")

In order to reduce falsely too long matches on some pages, we want to use [^\"]+? instead of .+?:

(href|HREF)=\"\K[^\"]+?\.png(?=\")

This disallows matches containing ” characters, hence preventing more than the tag being matched.

Usage example:

wget -qO- https://nasagrace.unl.edu/data/NASApublication/maps/ | grep -oP "(href|HREF)=\"\K[^\"]+?\.png(?=\")"

Output:

/data/NASApublication/maps/GRACE_SFSM_20201026.png
[...]

This systemd unit file autostarts your NodeJS service using npm start. Place it in /etc/systemd/system/myapplication.service (replace myapplication by the name of your application!)

[Unit]
Description=My application

[Service]
Type=simple
Restart=always
User=nobody
Group=nobody
WorkingDirectory=/opt/myapplication
ExecStart=/usr/bin/npm start

[Install]
WantedBy=multi-user.target

Replace:

• /opt/myapplication by the directory of your application (where package.json is located)
• User=nobody and Group=nobody by the user and group you want to run the service as
• Optionally you can add a custom decription instead of Description=My application

Then enable start at boot & start right now: (Remember to replace myapplication by the name of the service file you chose!)

sudo systemctl enable --now myapplication

Start by

sudo systemctl start myapplication

Restart by

sudo systemctl restart myapplication

Stop by

sudo systemctl stop myapplication

View & follow logs:

sudo journalctl -xfu myapplication

View logs in less:

sudo journalctl -xu myapplication

You can enable SSH in GRML using the ssh boot option. But if you have already started grub you can enable SSH using

Start ssh

Also remember to set a root password:

passwd

Run

mdadm --assemble --scan

and you’ll see all your MD devices.

Use this script (run as root) to automatically configure Linux to reboot after a Kernel hung event:

# Comment out previous kernel hung configuration
sed -i -e 's/kernel.hung_/#kernel.hung_/g' /etc/sysctl.conf
sed -i -e 's/kernel.panic/#kernel.panic/g' /etc/sysctl.conf
# Add configuration (which becomes effective after reboot)
cat <<EOF >>/etc/sysctl.conf

# Reboot after kernel hang
kernel.hung_task_timeout_secs=600
kernel.hung_task_panic=1
kernel.panic=300
EOF
# Activate config NOW
sysctl -w kernel.hung_task_timeout_secs=600
sysctl -w kernel.hung_task_panic=1
sysctl -w kernel.panic=300

The configuration works by converting the kernel hung event to a kernel panic event after kernel.hung_task_timeout_secs (10 minutes by default).

The configuration will be activated without reboot but it will also stay active after reboot.