Creating Linux Services with systemd
When writing web applications, I often need to offload computationally heavy tasks to an asynchronous worker script, schedule the task for later, or even write a daemon that listens on a socket to communicate directly with the client.
While there may sometimes be better tools for the job — always consider using existing software first, such as a task queue server — writing your own service gives you a level of flexibility you never get when tied to third-party software.
The cool thing is that creating a Linux service is fairly easy: write a long-running program in our favorite programming language, and then systemd
turn it into a service using .
program
Let's create a small service using PHP . I can see your eyebrows rising, but it works surprisingly well. We'll listen on UDP port 10000 and return any messages received via the ROT13 transformation:
<?php
$sock = socket_create(AF_INET, SOCK_DGRAM, SOL_UDP);
socket_bind($sock, '0.0.0.0', 10000);
echo "server started\n";
for (;;) {
socket_recvfrom($sock, $message, 1024, 0, $ip, $port);
$reply = str_rot13($message);
socket_sendto($sock, $reply, strlen($reply), 0, $ip, $port);
}
Let’s start it
$ php server.php
And test it in another terminal:
$ nc -u 127.0.0.1 10000
Hello, world!
Uryyb, jbeyq!
Cool, it works. Now we want this script to keep running, restart in case of failure (exit unexpectedly), and even survive a server reboot. This is systemd
where <script> comes into play.
Turn it into a service
Let's create a /etc/systemd/system/rot13.service
file called
[Unit]
Description=ROT13 demo service
After=network.target
StartLimitIntervalSec=0
[Service]
Type=simple
Restart=always
RestartSec=1
User=centos
ExecStart=/usr/bin/env php /path/to/server.php
[Install]
WantedBy=multi-user.target
we need to:
User=
Set your actual username afterExecStart=
Set the correct path to the script in
That's it. We can now start the service:
$ systemctl start rot13
And automatically have it start at boot time:
$ systemctl enable rot13
Deep understanding
Now that our service is (hopefully) working, it may be important to dive into the configuration options and ensure that it continues to work as expected.
Start in the right order
We might be wondering After=
what the directive does. It simply means that our service must be started after the network is ready. If the program wants the MySQL server to be up and running, it should add:
After=mysqld.service
Restart on exit
By default, systemd
the service will not be restarted if the program exits for any reason. For a service that must always be available, this is usually not what we want, so we instruct it to always restart on exit:
Restart=always
on-failure
We can also use just restart
if the exit status is not 0 .
By default, systemd
a restart is attempted after 100 milliseconds. We can specify the number of seconds to wait before attempting a restart using:
RestartSec=1
Avoiding pitfalls: start limits
I've personally run into this situation more than once. By default, when configured like ours is Restart=always
, systemd will give up on restarting our service if it starts more than 5 times within a 10 second interval. Forever.
There are two [Unit]
configuration options responsible for:
StartLimitBurst=5
StartLimitIntervalSec=10
RestartSec
The directive also has an impact on the results: if it is set to restart after 3 seconds, then we will never reach 5 failed retries in 10 seconds.
A simple fix that always works is to set StartLimitIntervalSec=0
. This way systemd will try to restart our service forever.
However, it is best to RestartSec
set it to at least 1 second to avoid putting too much stress on the server if problems occur.
As an alternative, we can keep the default settings and require systemd
a server restart when the startup limit is reached, using StartLimitAction=reboot
.
Is this really the case?
systemd
That's all it takes to create a Linux service
using : write a small configuration file that references our long-running program.
Systemd has been the default init system in RHEL/CentOS , Fedora , Ubuntu , Debian , and others for many years, so our server is likely ready to host Homebrew services!
For reprinting, please send an email to 1244347461@qq.com for approval. After obtaining the author's consent, kindly include the source as a link.
Related Articles
Restart PostgreSQL in Ubuntu 18.04
Publish Date:2025/04/09 Views:72 Category:PostgreSQL
-
This short article shows how to restart PostgreSQL in Ubuntu. Restart PostgreSQL Server in Ubuntu You can restart Postgres server in Ubuntu using the following command. Order: sudo service postgres restart Sometimes the above command does n
Issues to note when installing Apache on Linux
Publish Date:2025/04/08 Views:78 Category:OPERATING SYSTEM
-
As the most commonly used web server, Apache can be used in most computer operating systems. As a free and open source Unix-like operating system, Linux and Apache are a golden pair. This article will introduce the installation and use of A
How to decompress x.tar.xz format files under Linux
Publish Date:2025/04/08 Views:186 Category:OPERATING SYSTEM
-
A lot of software found today is in the tar.xz format, which is a lossless data compression file format that uses the LZMA compression algorithm. Like gzip and bzip2, it supports multiple file compression, but the convention is not to compr
Summary of vim common commands
Publish Date:2025/04/08 Views:115 Category:OPERATING SYSTEM
-
In Linux, the best editor should be vim. However, the complex commands behind vim's powerful functions also make us daunted. Of course, these commands do not need to be memorized by rote. As long as you practice using vim more, you can reme
Detailed explanation of command return value $? in Linux
Publish Date:2025/04/08 Views:58 Category:OPERATING SYSTEM
-
? is a special variable. This variable represents the return value of the previous command. That is to say, when we run certain commands, these commands will return a code after running. Generally, if the command is successfully run, the re
Common judgment formulas for Linux script shell
Publish Date:2025/04/08 Views:159 Category:OPERATING SYSTEM
-
In shell script programming, predicates are often used. There are two ways to use predicates, one is to use test, and the other is to use []. Let's take a look at how to use these two methods through two simple examples. Example 1 # test –
Shell script programming practice - specify a directory to delete files
Publish Date:2025/04/08 Views:98 Category:OPERATING SYSTEM
-
Usually, in Linux system we need to frequently delete some temporary files or junk files. If we delete them one by one manually, it will be quite troublesome. I have also been learning shell script programming recently, so I tried to write
Use of Linux command at - set time to execute command only once
Publish Date:2025/04/08 Views:158 Category:OPERATING SYSTEM
-
This article mainly involves a knowledge point, which is the atd service. Similar to this service is the crond service. The functions of these two services can be similar to the two functional functions of javascript. Those who have learned
Use of Linux command crontab - loop execution of set commands
Publish Date:2025/04/08 Views:170 Category:OPERATING SYSTEM
-
Compared with at , which executes a command only once, crontab, which we are going to talk about in this article, executes the set commands in a loop. Similarly, the use of crontab requires the support of the crond service. The service is s