Skip to content

services.md

Latest commit

 

History

History
1968 lines (1349 loc) · 36.9 KB

services.md

File metadata and controls

1968 lines (1349 loc) · 36.9 KB
Raw

Services, Daemons, and systemd

A service is a background program that provides a function to the system, users, or other programs.

Services usually run without direct user interaction. They may start automatically when the system boots, listen for network connections, process scheduled work, write logs, or respond to requests from other programs.

Common examples include:

Service Description
sshd SSH server for remote login
nginx Web server
httpd Apache web server
cron Scheduled task service
mariadb Database server
postgresql Database server
vsftpd FTP server
NetworkManager Network management service

The main idea is:

  • A service provides a function.
  • A daemon is often the background process that provides that function.
  • systemd manages services on many modern Linux systems.

Service Management Big Picture

Modern Linux systems commonly use systemd to manage services.

The command-line tool used to control systemd is:

systemctl

A simplified model looks like this:

+--------------------------------------------------+
|                Linux Operating System            |
|                                                  |
|    +------------------------+   +-------------+  |
|    |     Service Manager    |   | User        |  |
|    |       systemd          |<--| Commands    |  |
|    +------------------------+   | systemctl   |  |
|         |         ^             +-------------+  |
|         |         |                              |
|   start |         | stop/restart/status          |
| enable  |         | disable                      |
|         v         |                              |
|    +------------------------+                    |
|    |      Linux Service     |                    |
|    |                        |                    |
|    | - runs in background   |                    |
|    | - performs tasks       |                    |
|    | - listens on ports     |                    |
|    | - writes logs          |                    |
|    | - responds to events   |                    |
|    +------------------------+                    |
|                                                  |
+--------------------------------------------------+

The administrator uses systemctl to ask systemd to start, stop, enable, disable, restart, reload, or inspect services.

Services vs Processes

A process is a running instance of a program.

A service is a managed system function, usually represented by one or more processes.

For example:

Service:
sshd.service

Process:
PID 1234 running /usr/sbin/sshd

A service may have one process, several worker processes, or no long-running process if it is a one-shot task.

systemd unit: nginx.service
        |
        v
main process: nginx master
        |
        v
worker processes: nginx workers

So a service is not exactly the same thing as a process.

A service is the managed unit. A process is the actual running program.

Daemons

A daemon is a background process that usually runs independently of direct user control.

Many daemon names end with the letter d.

Examples:

  • sshd SSH daemon
  • httpd HTTP daemon
  • crond cron daemon
  • slapd LDAP daemon
  • vsftpd Very Secure FTP daemon

A daemon often starts during boot and keeps running until shutdown.

The word daemon describes how the program runs:

  • background
  • non-interactive
  • long-running
  • system-managed

Daemon vs Service

A daemon describes a background process.

A service describes a function provided to the system or users.

  • Daemon = how it runs
  • Service = what it provides

Example:

  • sshd is a daemon process.
  • sshd.service is the systemd service unit.
  • The service provides remote SSH login.

Daemons are often services, but not every service is a traditional daemon.

For example, a one-shot service may run once, complete a task, and exit.

systemd

systemd is a service manager and initialization system used by many modern Linux distributions.

It is responsible for:

  • starting system services
  • stopping services during shutdown
  • tracking service state
  • managing dependencies
  • starting services in the correct order
  • capturing service logs through journald
  • handling sockets, timers, mounts, and targets

At boot time, systemd starts the services and units needed to bring the system into a usable state.

Units

In systemd, a unit is something systemd can manage.

Common unit types include:

  • .service system service
  • .socket socket activation unit
  • .target group of units, similar to a boot milestone
  • .timer scheduled activation unit
  • .mount mount point
  • .automount automatic mount point
  • .path path-based activation
  • .device device unit

A service unit usually ends in:

.service

Example:

  • sshd.service
  • nginx.service
  • vsftpd.service
  • postgresql.service

To list active units:

systemctl list-units

To list all service units, including inactive ones:

systemctl list-units --type=service --all

Service States

A service can have several important states.

  • loaded systemd successfully read the unit file
  • active the unit is currently active
  • inactive the unit is not active
  • failed the unit tried to start but failed
  • enabled the unit is configured to start at boot
  • disabled the unit is not configured to start at boot
  • static the unit cannot be enabled directly
  • masked the unit is blocked from being started

Important distinction:

  • start/stop controls the current running state.
  • enable/disable controls boot-time behavior.

For example:

  • A service can be active but disabled.

  • That means it is running now, but will not automatically start at boot.

  • A service can be inactive but enabled.

  • That means it is stopped now, but should start at boot.

Checking Service Status

Use:

systemctl status sshd.service

or sometimes, depending on distribution:

systemctl status ssh.service

Example output:

● sshd.service - OpenSSH server daemon
     Loaded: loaded (/usr/lib/systemd/system/sshd.service; enabled; preset: enabled)
     Active: active (running) since Mon 2026-06-01 10:30:00 CEST; 5min ago
   Main PID: 1234 (sshd)
      Tasks: 1
     Memory: 6.2M
        CPU: 120ms
     CGroup: /system.slice/sshd.service
             └─1234 /usr/sbin/sshd -D

Interpretation:

  • Loaded means systemd found and loaded the unit file.
  • enabled means the service is configured to start at boot.
  • active (running) means it is currently running.
  • Main PID shows the main service process.

If a service is broken, status output often includes recent log lines.

Starting and Stopping Services

To start a service immediately:

sudo systemctl start sshd.service

To stop it immediately:

sudo systemctl stop sshd.service

To restart it:

sudo systemctl restart sshd.service

To reload configuration without fully stopping the service, if supported:

sudo systemctl reload sshd.service

If you are not sure whether reload is supported, use:

sudo systemctl reload-or-restart sshd.service

Enabling and Disabling Services

To enable a service at boot:

sudo systemctl enable sshd.service

Example output:

Created symlink /etc/systemd/system/multi-user.target.wants/sshd.service → /usr/lib/systemd/system/sshd.service.

Interpretation:

  • systemd created a startup link.
  • The service should start automatically at boot.

To disable a service at boot:

sudo systemctl disable sshd.service

This prevents automatic startup, but it does not necessarily stop the service right now.

To enable and start at the same time:

sudo systemctl enable --now sshd.service

To disable and stop at the same time:

sudo systemctl disable --now sshd.service

Masking Services

Masking is stronger than disabling.

A disabled service can still be started manually.

A masked service cannot be started unless it is unmasked first.

Mask a service:

sudo systemctl mask service-name.service

Unmask it:

sudo systemctl unmask service-name.service

Use masking carefully. It is useful when you want to prevent a service from being started by mistake or by another dependency.

Listing Services

List active services:

systemctl list-units --type=service

List all services:

systemctl list-units --type=service --all

List enabled services:

systemctl list-unit-files --type=service | grep enabled

List failed services:

systemctl --failed

Example failed output:

UNIT                  LOAD   ACTIVE SUB    DESCRIPTION
myapp.service          loaded failed failed My Test Service

Interpretation:

  • myapp.service tried to run but failed.
  • Use systemctl status and journalctl to inspect why.

Service Logs with journalctl

systemd services usually log to the system journal.

To view logs for a service:

journalctl -u sshd.service

To view logs from the current boot:

journalctl -u sshd.service -b

To follow logs live:

journalctl -u sshd.service -f

To show recent lines:

journalctl -u sshd.service -n 50

Example output:

Jun 01 10:30:00 host systemd[1]: Started OpenSSH server daemon.
Jun 01 10:32:10 host sshd[1500]: Accepted publickey for adam from 192.168.1.20 port 53022

Interpretation:

  • The service started successfully.
  • A user logged in using an SSH public key.

Dependencies

Services may depend on other units.

For example, a web service may need the network to be online before it starts.

To list dependencies:

systemctl list-dependencies nginx.service

To see what depends on a service:

systemctl list-dependencies --reverse nginx.service

To check dependencies of a target:

systemctl list-dependencies multi-user.target

Example:

systemctl list-dependencies multi-user.target | grep sshd

If output appears, sshd.service is part of that dependency tree.

Targets

A target is a group of units.

Targets are similar to boot milestones.

Common targets:

  • basic.target basic system initialization
  • multi-user.target multi-user text/server mode
  • graphical.target graphical desktop mode
  • rescue.target rescue mode
  • emergency.target minimal emergency shell

To see the default target:

systemctl get-default

Example output:

graphical.target

Set default target:

sudo systemctl set-default multi-user.target

Older Service Management Tools

Older Linux systems may use SysV init scripts.

Common tools include:

  • service
  • chkconfig
  • /etc/init.d/

List old init scripts:

ls /etc/init.d

Start a service using the older interface:

sudo service httpd start

Enable a service on older Red Hat-style systems:

sudo chkconfig httpd on

Modern systems often keep the service command as a compatibility layer, but systemctl is preferred on systemd systems.

Installing and Managing an FTP Server Example

The uploaded notes use vsftpd as an example service.

FTP is useful for learning service management, ports, logs, and testing, but it is usually not preferred for secure production file transfer unless configured carefully with encryption. For secure remote administration, SSH/SFTP is usually safer.

Installing vsftpd

On Red Hat/CentOS-style systems:

sudo yum install vsftpd

or newer systems:

sudo dnf install vsftpd

On Debian/Ubuntu systems:

sudo apt install vsftpd

Check if it is installed on RPM-based systems:

rpm -qa | grep vsftpd

Example output:

vsftpd-3.0.2-29.el7.x86_64

Interpretation:

The vsftpd package is installed.

Configuring vsftpd

The configuration file is commonly:

/etc/vsftpd/vsftpd.conf

Open it:

sudo vi /etc/vsftpd/vsftpd.conf

Common options:

  • anonymous_enable=NO
  • local_enable=YES
  • write_enable=YES

Meaning:

  • anonymous_enable controls anonymous FTP access.
  • local_enable allows local Linux users to log in.
  • write_enable allows uploads and file changes.

After changing configuration, restart the service:

sudo systemctl restart vsftpd.service

Starting and Enabling vsftpd

Start now:

sudo systemctl start vsftpd.service

Enable at boot:

sudo systemctl enable vsftpd.service

Enable and start together:

sudo systemctl enable --now vsftpd.service

Check status:

systemctl status vsftpd.service

Example output:

● vsftpd.service - Vsftpd ftp daemon
   Loaded: loaded (/usr/lib/systemd/system/vsftpd.service; enabled; vendor preset: disabled)
   Active: active (running) since Mon 2026-06-01 12:30:50 CEST; 5min ago

Interpretation:

  • The service is installed.
  • It is enabled for boot.
  • It is currently running.

Checking Listening Ports

FTP normally listens on TCP port 21.

Use ss:

sudo ss -tulnp | grep ':21'

Example output:

tcp LISTEN 0 32 0.0.0.0:21 0.0.0.0:* users:(("vsftpd",pid=1234,fd=3))

Interpretation:

  • vsftpd is listening on TCP port 21.
  • The process ID is 1234.
  • The service is ready for incoming FTP connections.

Older command:

sudo netstat -tulnp | grep ftp

On modern Linux, ss is usually preferred.

Testing an FTP Service

Test locally:

ftp localhost

Example output:

Connected to localhost (127.0.0.1).
220 (vsFTPd 3.0.2)
Name (localhost:user):

Interpretation:

  • The FTP server accepted a connection.
  • The 220 message means the server is ready.

Test from another machine:

ftp SERVER_IP

If local testing works but remote testing fails, check:

  • firewall rules
  • network connectivity
  • service listen address
  • SELinux or security policy
  • router or NAT rules

Checking FTP Logs

Common log file:

sudo tail -f /var/log/vsftpd.log

Example output:

Mon Jun  1 12:35:15 2026 [pid 1235] CONNECT: Client "::1"
Mon Jun  1 12:35:16 2026 [pid 1235] [anonymous] OK LOGIN: Client "::1"
Mon Jun  1 12:36:20 2026 [pid 1236] [user] UPLOAD: Client "::1", "/home/user/testfile.txt"

Interpretation:

  • A client connected.
  • Anonymous login succeeded.
  • A file upload occurred.

You can also use systemd logs:

journalctl -u vsftpd.service -f

Creating a Custom systemd Service

A custom service is defined using a unit file.

System-level service unit files are usually stored in:

/etc/systemd/system/

A simple service file has three main sections:

  • [Unit]
  • [Service]
  • [Install]

Unit File Sections

[Unit] describes the service and its ordering or dependencies.

Common directives:

  • Description=
  • Documentation=
  • After=
  • Wants=
  • Requires=

[Service] describes how the service runs.

Common directives:

  • Type=
  • ExecStart=
  • ExecStop=
  • ExecReload=
  • Restart=
  • User=
  • WorkingDirectory=
  • Environment=

[Install] describes how the service is enabled at boot.

Common directive:

WantedBy=multi-user.target

Simple Custom Service Example

Create an executable script:

sudo tee /usr/local/bin/sample-service.sh > /dev/null <<'EOF'
#!/bin/bash
while true; do
    echo "$(date): sample service is running"
    sleep 10
done
EOF

sudo chmod +x /usr/local/bin/sample-service.sh

Create a service unit:

sudo tee /etc/systemd/system/sample.service > /dev/null <<'EOF'
[Unit]
Description=Sample Script Service

[Service]
Type=simple
ExecStart=/usr/local/bin/sample-service.sh
Restart=always

[Install]
WantedBy=multi-user.target
EOF

Reload systemd:

sudo systemctl daemon-reload

Start the service:

sudo systemctl start sample.service

Check status:

systemctl status sample.service

Example output:

● sample.service - Sample Script Service
     Loaded: loaded (/etc/systemd/system/sample.service; disabled)
     Active: active (running) since Mon 2026-06-01 13:00:00 CEST
   Main PID: 2401 (sample-service)

Interpretation:

  • The custom service is loaded.
  • It is currently running.
  • It is disabled, so it will not start at boot unless enabled.

Enable it:

sudo systemctl enable sample.service

Common Service Types

The Type= field controls how systemd decides whether a service has started successfully.

Common types:

  • simple process started by ExecStart is the main process
  • exec similar to simple, but waits until exec succeeds
  • forking service forks into background
  • oneshot command runs once and exits
  • notify service tells systemd when it is ready
  • idle waits until other startup jobs are dispatched

For most simple scripts, use:

Type=simple

For one-time tasks, use:

Type=oneshot

Scenario 1: Simulate a Stopped Service

Practice identifying when a service is installed but not running.

Simulate the Problem

Use the sample service or another safe test service:

sudo systemctl stop sample.service

Check Status

systemctl status sample.service

Example output:

● sample.service - Sample Script Service
     Loaded: loaded (/etc/systemd/system/sample.service; enabled)
     Active: inactive (dead)

Interpretation:

  • The service exists.
  • It is enabled for boot.
  • It is not currently running.
  • This is a runtime state problem, not a boot enablement problem.

Fix

sudo systemctl start sample.service

Verify:

systemctl status sample.service

Expected:

Active: active (running)

Scenario 2: Simulate a Disabled Service

Understand the difference between stopped and disabled.

Simulate the Problem

sudo systemctl disable sample.service

Check:

systemctl is-enabled sample.service

Example output:

disabled

The service may still be running.

Check:

systemctl is-active sample.service

Example output:

active

Interpretation:

  • The service is running now.
  • But it will not automatically start at boot.
  • This is boot-time configuration, not current runtime state.

Fix

sudo systemctl enable sample.service

Or enable and start:

sudo systemctl enable --now sample.service

Scenario 3: Simulate a Service That Fails Because ExecStart Is Wrong

Practice diagnosing a failed custom service.

Simulate the Problem

Create a broken unit:

sudo tee /etc/systemd/system/broken.service > /dev/null <<'EOF'
[Unit]
Description=Broken Service Example

[Service]
Type=simple
ExecStart=/not/a/real/path

[Install]
WantedBy=multi-user.target
EOF

Reload and start:

sudo systemctl daemon-reload
sudo systemctl start broken.service

Check Status

systemctl status broken.service

Example output:

● broken.service - Broken Service Example
     Loaded: loaded (/etc/systemd/system/broken.service; disabled)
     Active: failed (Result: exit-code)
    Process: 2500 ExecStart=/not/a/real/path (code=exited, status=203/EXEC)

Interpretation:

  • status=203/EXEC usually means systemd could not execute the command.
  • The ExecStart path is wrong, missing, or not executable.

Check Logs

journalctl -u broken.service -b

Example output:

systemd[1]: broken.service: Failed to locate executable /not/a/real/path
systemd[1]: broken.service: Failed at step EXEC spawning /not/a/real/path: No such file or directory

Fix

Edit the unit:

sudo systemctl edit --full broken.service

Set a valid executable path.

Then reload and restart:

sudo systemctl daemon-reload
sudo systemctl restart broken.service

Scenario 4: Simulate a Service Crash and Auto-Restart

Show how Restart= affects service behavior.

Create a Crashing Script

sudo tee /usr/local/bin/crash-service.sh > /dev/null <<'EOF'
#!/bin/bash
echo "crash service started"
sleep 2
echo "crash service exiting with error"
exit 1
EOF

sudo chmod +x /usr/local/bin/crash-service.sh

Create the unit:

sudo tee /etc/systemd/system/crash.service > /dev/null <<'EOF'
[Unit]
Description=Crash Test Service

[Service]
Type=simple
ExecStart=/usr/local/bin/crash-service.sh
Restart=always
RestartSec=3

[Install]
WantedBy=multi-user.target
EOF

Start it:

sudo systemctl daemon-reload
sudo systemctl start crash.service

Check Status

systemctl status crash.service

Example output:

● crash.service - Crash Test Service
     Loaded: loaded (/etc/systemd/system/crash.service; disabled)
     Active: activating (auto-restart) (Result: exit-code)
    Process: 2701 ExecStart=/usr/local/bin/crash-service.sh (code=exited, status=1/FAILURE)

Interpretation:

  • The service exits with an error.
  • systemd is restarting it because Restart=always is set.
  • This can create a restart loop if the underlying problem is not fixed.

Check Logs

journalctl -u crash.service -f

Example output:

crash-service.sh[2701]: crash service started
crash-service.sh[2701]: crash service exiting with error
systemd[1]: crash.service: Main process exited, code=exited, status=1/FAILURE
systemd[1]: crash.service: Scheduled restart job, restart counter is at 3.

Fix

Fix the script or change restart policy.

Stop the test:

sudo systemctl stop crash.service

Scenario 5: Simulate a Port Conflict

Understand why a network service may fail to start when its port is already in use.

Simulate the Problem

Start a temporary listener on port 8080:

python3 -m http.server 8080

In another terminal, check the port:

ss -tulnp | grep ':8080'

Example output:

tcp LISTEN 0 5 0.0.0.0:8080 0.0.0.0:* users:(("python3",pid=3001,fd=3))

Now start another service configured to use the same port.

Example failure log may look like:

Address already in use
Failed to bind to 0.0.0.0:8080

Interpretation:

  • The service is not failing because systemd is broken.
  • It is failing because another process already owns the port.

Fix

Find and stop the conflicting process:

sudo ss -tulnp | grep ':8080'
sudo kill 3001

Or configure one service to use a different port.

Scenario 6: Simulate a Service Configuration Error

Practice checking service logs after a bad configuration change.

Simulate the Problem

For a real service like nginx, a bad config might cause restart failure.

For a safer custom example, create a script that exits when a config file contains bad.

sudo tee /usr/local/bin/config-check-service.sh > /dev/null <<'EOF'
#!/bin/bash
if grep -q bad /etc/config-check-service.conf; then
    echo "Bad configuration found"
    exit 1
fi

echo "Configuration OK"
sleep infinity
EOF

sudo chmod +x /usr/local/bin/config-check-service.sh
echo "bad=true" | sudo tee /etc/config-check-service.conf

Create service:

sudo tee /etc/systemd/system/config-check.service > /dev/null <<'EOF'
[Unit]
Description=Config Check Service

[Service]
Type=simple
ExecStart=/usr/local/bin/config-check-service.sh

[Install]
WantedBy=multi-user.target
EOF

Start it:

sudo systemctl daemon-reload
sudo systemctl start config-check.service

Check Status

systemctl status config-check.service

Example output:

Active: failed (Result: exit-code)
Process: 3100 ExecStart=/usr/local/bin/config-check-service.sh (code=exited, status=1/FAILURE)

Check Logs

journalctl -u config-check.service -b

Example output:

config-check-service.sh[3100]: Bad configuration found

Interpretation:

  • The service started successfully enough to run the script.
  • The script rejected the configuration and exited.
  • The issue is application configuration, not systemd itself.

Fix

echo "good=true" | sudo tee /etc/config-check-service.conf
sudo systemctl restart config-check.service

Scenario 7: Simulate a Missing Dependency or Wrong Startup Order

Understand service ordering with After= and Wants=.

Suppose a service needs the network before starting.

A unit may include:

[Unit]
Description=Network Dependent App
Wants=network-online.target
After=network-online.target

Meaning:

  • Wants asks systemd to start network-online.target too.
  • After controls ordering, so this service starts after network-online.target.

Check Dependencies

systemctl list-dependencies network-online.target

Check reverse dependencies:

systemctl list-dependencies --reverse network-online.target

Interpretation:

  • If a network-dependent service starts too early, it may fail because DNS or network routes are not ready.
  • Use After= and Wants= carefully to express startup requirements.

Important note:

  • After= only controls order.
  • It does not automatically pull in the dependency.
  • Use Wants= or Requires= when the other unit should also be started.

Scenario 8: Simulate a Failed Service and Reset Its State

Understand failed state and how to clear it.

Simulate Failure

Use the broken service:

sudo systemctl start broken.service

Check failed services:

systemctl --failed

Example output:

UNIT             LOAD   ACTIVE SUB    DESCRIPTION
broken.service   loaded failed failed Broken Service Example

Interpretation:

  • systemd remembers that the service failed.
  • Even after fixing the issue, the failed state may remain until reset.

Reset Failed State

sudo systemctl reset-failed broken.service

Check again:

systemctl --failed

Expected output:

0 loaded units listed.

Scenario 9: Simulate a Masked Service

Understand why a service may refuse to start even though it exists.

Simulate

sudo systemctl mask sample.service

Try to start it:

sudo systemctl start sample.service

Example output:

Failed to start sample.service: Unit sample.service is masked.

Interpretation:

  • The service is intentionally blocked.
  • Starting it is not allowed while it is masked.

Fix

sudo systemctl unmask sample.service
sudo systemctl start sample.service

Scenario 10: Simulate a Service Consuming Too Much CPU

Use service monitoring tools to identify a resource-heavy service.

Create a CPU-Heavy Service

sudo tee /usr/local/bin/cpu-service.sh > /dev/null <<'EOF'
#!/bin/bash
while true; do
    :
done
EOF

sudo chmod +x /usr/local/bin/cpu-service.sh

Create unit:

sudo tee /etc/systemd/system/cpu-test.service > /dev/null <<'EOF'
[Unit]
Description=CPU Test Service

[Service]
Type=simple
ExecStart=/usr/local/bin/cpu-service.sh

[Install]
WantedBy=multi-user.target
EOF

Start:

sudo systemctl daemon-reload
sudo systemctl start cpu-test.service

Check with systemctl status

systemctl status cpu-test.service

Example output:

● cpu-test.service - CPU Test Service
     Active: active (running)
   Main PID: 3500 (cpu-service.sh)
      Tasks: 1
        CPU: 1min 20s

Check with top

top -p 3500

Example output:

PID USER  PR NI  VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3500 root 20  0  4000 900 800 R 99.9  0.0  1:20 cpu-service.sh

Interpretation:

  • The service is active.
  • Its main process is consuming nearly one full CPU core.
  • This is a service-level CPU bottleneck.

Fix

Stop the test service:

sudo systemctl stop cpu-test.service

For real services, investigate configuration, loops, workload spikes, or bugs.

Scenario 11: Create a One-Shot Service

Understand services that run once and exit successfully.

Create script:

sudo tee /usr/local/bin/oneshot-task.sh > /dev/null <<'EOF'
#!/bin/bash
echo "$(date): oneshot task ran" >> /var/log/oneshot-task.log
EOF

sudo chmod +x /usr/local/bin/oneshot-task.sh

Create unit:

sudo tee /etc/systemd/system/oneshot-task.service > /dev/null <<'EOF'
[Unit]
Description=One Shot Task Example

[Service]
Type=oneshot
ExecStart=/usr/local/bin/oneshot-task.sh

[Install]
WantedBy=multi-user.target
EOF

Run:

sudo systemctl daemon-reload
sudo systemctl start oneshot-task.service

Check status:

systemctl status oneshot-task.service

Example output:

Active: inactive (dead)
Process: 3600 ExecStart=/usr/local/bin/oneshot-task.sh (code=exited, status=0/SUCCESS)

Interpretation:

  • The service is inactive now because the task finished.
  • This is normal for Type=oneshot.
  • status=0/SUCCESS means it completed successfully.

Scenario 12: Timers as Scheduled Services

Understand systemd timers as an alternative to cron.

A timer activates a service on a schedule.

Create a timer for the one-shot task:

sudo tee /etc/systemd/system/oneshot-task.timer > /dev/null <<'EOF'
[Unit]
Description=Run oneshot task every minute

[Timer]
OnCalendar=*:0/1
Persistent=true

[Install]
WantedBy=timers.target
EOF

Reload and start timer:

sudo systemctl daemon-reload
sudo systemctl enable --now oneshot-task.timer

List timers:

systemctl list-timers --all

Example output:

NEXT                        LEFT LAST                        PASSED UNIT                 ACTIVATES
Mon 2026-06-01 14:01:00     30s  Mon 2026-06-01 14:00:00     30s    oneshot-task.timer   oneshot-task.service

Interpretation:

  • The timer runs oneshot-task.service every minute.
  • The ACTIVATES column shows which service the timer starts.

Service Troubleshooting Workflow

When a service is not working, troubleshoot in layers.

  1. Is the service installed?
  2. Does systemd know about the unit?
  3. Is it enabled for boot?
  4. Is it active right now?
  5. Did it fail?
  6. What do the logs say?
  7. Is the configuration valid?
  8. Are required ports available?
  9. Are dependencies ready?
  10. Are permissions, users, or paths wrong?

Step 1: Check Unit Exists

systemctl status service-name.service

If output says:

Unit service-name.service could not be found.

then the package may not be installed, or the unit name may be wrong.

Search unit files:

systemctl list-unit-files | grep service-name

Step 2: Check Active State

systemctl is-active service-name.service

Possible output:

  • active
  • inactive
  • failed
  • unknown

Step 3: Check Boot State

systemctl is-enabled service-name.service

Possible output:

  • enabled
  • disabled
  • static
  • masked

Step 4: Read Status

systemctl status service-name.service

Look for:

  • Loaded
  • Active
  • Main PID
  • status code
  • recent log lines

Step 5: Read Logs

journalctl -u service-name.service -b

Follow logs:

journalctl -u service-name.service -f

Step 6: Check Ports

For network services:

sudo ss -tulnp

Example:

sudo ss -tulnp | grep ':80'

This helps answer:

  • Is the service listening?
  • Is another process already using the port?
  • Is it listening on localhost only or all interfaces?

Step 7: Check Processes

ps -ef | grep service-name

or use systemd’s process tree:

systemctl status service-name.service

Step 8: Check Configuration

Many services have a built-in config test.

Examples:

sudo nginx -t
sudo apachectl configtest
sudo sshd -t

If available, always test configuration before restarting production services.

Common Service Problems and Fixes

Problem: Service Not Found

Example:

Unit nginx.service could not be found.

Possible causes:

  • package not installed
  • wrong service name
  • unit file not reloaded
  • custom unit placed in wrong location

Fix:

sudo systemctl daemon-reload
systemctl list-unit-files | grep nginx

Problem: Service Failed to Start

Check:

systemctl status service-name.service
journalctl -u service-name.service -b

Common causes:

  • bad configuration
  • missing executable
  • wrong permissions
  • port already in use
  • missing dependency
  • invalid user or group
  • missing directory

Problem: Service Running but Not Reachable

Check:

systemctl status service-name.service
sudo ss -tulnp
journalctl -u service-name.service -b

Possible causes:

  • service listens only on localhost
  • firewall blocks access
  • wrong port
  • network problem
  • application-level access control
  • SELinux/AppArmor policy

Problem: Service Starts Manually but Not at Boot

Check:

systemctl is-enabled service-name.service
systemctl status service-name.service

If disabled:

sudo systemctl enable service-name.service

If enabled but still fails at boot, check:

journalctl -u service-name.service -b
systemctl list-dependencies service-name.service

Possible causes:

  • starts before network is ready
  • missing mount at boot
  • environment differs at boot
  • dependency missing
  • permissions or path issue

Problem: Service Keeps Restarting

Check:

systemctl status service-name.service
journalctl -u service-name.service -b

Look for:

  • Scheduled restart job
  • restart counter
  • exit-code
  • core dump
  • configuration errors

Possible causes:

  • application crashes
  • bad config
  • missing file
  • health check failure
  • Restart=always hides repeated failure

Useful Command Summary

Service state:

systemctl status service.service
systemctl is-active service.service
systemctl is-enabled service.service
systemctl --failed

Start and stop:

sudo systemctl start service.service
sudo systemctl stop service.service
sudo systemctl restart service.service
sudo systemctl reload service.service
sudo systemctl reload-or-restart service.service

Boot behavior:

sudo systemctl enable service.service
sudo systemctl disable service.service
sudo systemctl enable --now service.service
sudo systemctl disable --now service.service

Masking:

sudo systemctl mask service.service
sudo systemctl unmask service.service

Listing:

systemctl list-units --type=service
systemctl list-units --type=service --all
systemctl list-unit-files --type=service
systemctl list-timers --all

Logs:

journalctl -u service.service
journalctl -u service.service -b
journalctl -u service.service -f
journalctl -u service.service -n 50

Dependencies:

systemctl list-dependencies service.service
systemctl list-dependencies --reverse service.service

Ports and processes:

sudo ss -tulnp
ps -ef | grep service-name

Custom units:

sudo systemctl daemon-reload
sudo systemctl edit --full service.service
sudo systemctl reset-failed service.service

Safe Lab Cleanup

If you created the sample services, clean them up:

sudo systemctl stop sample.service broken.service crash.service config-check.service cpu-test.service oneshot-task.timer oneshot-task.service 2>/dev/null

sudo systemctl disable sample.service broken.service crash.service config-check.service cpu-test.service oneshot-task.timer oneshot-task.service 2>/dev/null

sudo rm -f /etc/systemd/system/sample.service
sudo rm -f /etc/systemd/system/broken.service
sudo rm -f /etc/systemd/system/crash.service
sudo rm -f /etc/systemd/system/config-check.service
sudo rm -f /etc/systemd/system/cpu-test.service
sudo rm -f /etc/systemd/system/oneshot-task.service
sudo rm -f /etc/systemd/system/oneshot-task.timer

sudo rm -f /usr/local/bin/sample-service.sh
sudo rm -f /usr/local/bin/crash-service.sh
sudo rm -f /usr/local/bin/config-check-service.sh
sudo rm -f /usr/local/bin/cpu-service.sh
sudo rm -f /usr/local/bin/oneshot-task.sh
sudo rm -f /etc/config-check-service.conf

sudo systemctl daemon-reload
sudo systemctl reset-failed

Challenges

  1. Use systemctl list-units --type=service --all to list active and inactive services. Pick five services and explain what each one does.
  2. Choose one service and inspect it with systemctl status, systemctl is-active, and systemctl is-enabled. Explain the difference between active and enabled.
  3. Start, stop, restart, enable, and disable a safe test service. Record the command and output for each action.
  4. Use journalctl -u service-name -b to inspect logs for a service. Identify when it started and whether it produced errors.
  5. Install and configure a simple service such as vsftpd, nginx, or openssh-server on a lab machine. Verify that it is running and listening on the expected port.
  6. Use ss -tulnp to identify which services are listening on network ports.
  7. Create a custom systemd service that runs a script. Start it, check its status, view logs, enable it, and then clean it up.
  8. Create a broken service with an invalid ExecStart. Start it, inspect the failure, fix it, and restart it.
  9. Create a one-shot service and a timer that runs it every minute. Use systemctl list-timers --all to verify the schedule.
  10. Explain the difference between a process, daemon, and service using examples such as sshd, nginx, and a one-shot backup task.