Periodic processes

Table of contents

• 1. Text
  • 1.1. Python script
  • 1.2. Service
  • 1.3. Timer
• 2. Hints
  • 2.1. System reference manuals
  • 2.2. Python script
    • 2.2.1. Program execution
    • 2.2.2. Interactive mode
    • 2.2.3. Documentation
    • 2.2.4. Program structure
    • 2.2.5. Recursion
    • 2.2.6. String matching
  • 2.3. Service
    • 2.3.1. Unit file structure
    • 2.3.2. Unit file location
    • 2.3.3. Service management
    • 2.3.4. Logs
  • 2.4. Timer
    • 2.4.1. Unit file structure
    • 2.4.2. Unit file location
    • 2.4.3. Timer verification
    • 2.4.4. Timer management
• 3. Solution
• Licenses

1. Text

1.1. Python script

Write a Python script that analyzes log files in a specified directory and its subdirectories, extracting lines that contain a specified string and writing them to output files in a backup directory named backup, created (if necessary) in ~. In your home directory, create a directory called log-extractor and, within it, a file named app.py, using this template:

# first and last name:
# student id:
#
# path:
 
import argparse
import os
import sys
 
def main():
    pass
 
 
if __name__ == "__main__":
    main()

The script must accept exactly two command-line arguments, parsed using the argparse module. The first argument, --path, is a mandatory string specifying the absolute path of the directory to analyze. The second argument, --pattern, is a mandatory string specifying the pattern to search for within log files, i.e., those files with a .log extension.

After parsing, validate both inputs: check that the path is absolute (os.path.isabs), exists (os.path.exists), and is a directory (os.path.isdir); verify that --pattern is a non-empty string. If any check fails, print an explanatory error message to standard error (print) and exit with a non-zero status code (sys.exit).

After validation, ensure the backup directory exists in ~ (os.path.expanduser, os.makedirs). Recursively traverse the directory tree at the provided path (os.listdir, os.path.join, os.path.isdir). For each log file encountered (os.path.isfile, str.endswith), check if it contains the specified pattern by reading its content (open). If the pattern is found, extract the corresponding lines and write them to a file with the same name in the ~/backup directory (open, file.writelines), printing a log message to standard output (print) indicating the number of lines written and the destination file name.

For example, running:

$ python ~/log-extractor/app.py \
    --path ~/logs \
    --pattern ERROR

the script will extract all lines containing the string ERROR from files with a .log extension in ~/logs and its subdirectories, writing them to corresponding files in the ~/backup folder.

1.2. Service

Create a service unit named log-extractor.service in your user instance of systemd. Configure it to run ~/log-extractor/app.py with the arguments --path %h/logs and --pattern ERROR. Use this template:

# first and last name:
# student id:
#
# path:

1.3. Timer

Create a timer unit named log-extractor.timer in your user instance of systemd. Configure it to activate log-extractor.service every Monday and Friday at 02:00. Use this template:

# first and last name:
# student id:
#
# path:
#
# command to enable the timer:
# command to start the timer:

2. Hints

2.1. System reference manuals

man displays the system reference manuals. For example, man:systemctl(1) translates to

$ man 1 systemctl

2.2. Python script

2.2.1. Program execution

To run a Python program

$ python3 <your_program>.py

Warning

Depending on your system, you may have to use python or python3.

2.2.2. Interactive mode

The interactive mode starts a read-eval-print loop (REPL) where you can type Python expressions and see their results immediately. This is useful for experimenting with functions and modules before writing a full program.

To enable the interactive mode

$ python3

2.2.3. Documentation

help() displays documentation for modules, functions, classes, and keywords in the Python interactive interpreter. For example, to read the documentation for the argparse module

$ python3
>>> import argparse
>>> help(argparse)

To get help on a specific function

$ python3
>>> import os.path
>>> help(os.path.isabs)

2.2.4. Program structure

The following extends the template from §1.1 with comments that map to each step described in the text.

# first name and last name:
# student id:
#
# path:
 
import argparse
import os
import sys
 
 
def find(filename, pattern):
    # read file, return matching lines
    pass
 
 
def dump(filename, lines):
    # write lines to file
    pass
 
 
def walk(log_dir, backup_dir, pattern):
    # recursive traversal
    pass
 
 
def main():
    # argument parsing
 
    # validation
 
    # ensure backup directory exists
 
    # scan
    pass
 
 
if __name__ == "__main__":
    main()

Tip

os.makedirs raises FileExistsError if the directory already exists. Pass exist_ok=True to silently skip creation when the directory is already there.

2.2.5. Recursion

The following example recursively lists all file paths in a directory tree.

import os
 
 
def walk(basepath):
    for filename in os.listdir(basepath):
        path = os.path.join(basepath, filename)
        if os.path.isfile(path):
            print(path)
        elif os.path.isdir(path):
            walk(path)

2.2.6. String matching

The following example reads all lines from a file and keeps only those that contain a given string.

with open("server.log", "r") as f:
    lines = f.readlines()
 
matching = [line for line in lines if "timeout" in line]

2.3. Service

2.3.1. Unit file structure

The following extends the template from §1.2 with comments that map to each section of the unit file.

# first name and last name:
# student id:
#
# path:
 
[Unit]
# human-readable description
 
[Service]
# how to run the service

An example of a service unit file is provided here. The available directives are documented in man:systemd.unit(5) (for [Unit]) and man:systemd.service(5) (for [Service]).

Tip

%h expands to the home directory of the user running the service. For example, WorkingDirectory=%h/log-extractor expands to WorkingDirectory=/home/ubuntu/log-extractor.

Tip

Add Environment=PYTHONUNBUFFERED=1 to [Service] so that print output shows up in journalctl right away.

Warning

This service unit file does not have an [Install] section. The timer unit is responsible for activating the service, so the service does not need to be enabled or started on boot independently.

2.3.2. Unit file location

The directories where systemd reads unit files are documented in man:systemd.unit(5). The recommended directory for user units is ~/.config/systemd/user. If it does not exist, create it

$ mkdir -p ~/.config/systemd/user

Then copy the unit file there

$ cp ~/log-extractor/log-extractor.service ~/.config/systemd/user

2.3.3. Service management

The commands to control systemd are listed here. Append the --user option to control the user instance instead of the system-wide one. See also man:systemctl(1).

After copying the unit file:

• List unit files to verify that systemd sees the new file (list-unit-files)
• Inspect the unit file to verify its content (cat)
• Reload unit files so that systemd picks up any changes (daemon-reload)
• Start the service (start)
• Check the status to confirm it ran correctly (status)

2.3.4. Logs

journalctl queries the systemd journal. Append the --user option to read logs from the user instance. For example

$ journalctl --user -u log-extractor.service

See also man:journalctl(1).

2.4. Timer

2.4.1. Unit file structure

The following extends the template from §1.3 with comments that map to each section of the unit file.

# first name and last name:
# student id:
#
# path:
#
# command to enable the timer:
# command to start the timer:
 
[Unit]
# human-readable description
 
[Timer]
# which service to activate
# schedule (calendar expression)
 
[Install]
# start on boot

The available directives are documented in man:systemd.timer(5) (for [Timer]) and man:systemd.unit(5) (for [Unit] and [Install]). The timer types and calendar expressions are covered here.

Tip

Use WantedBy=timers.target in the [Install] section. timers.target is a dedicated target that collects all timer units and is activated early during boot. This is preferable to default.target, which is about bringing the system to its normal operating state, not about managing timers.

2.4.2. Unit file location

Timer unit files go in the same directory as service unit files: ~/.config/systemd/user.

$ cp ~/log-extractor/log-extractor.timer ~/.config/systemd/user

2.4.3. Timer verification

systemd-analyze calendar parses calendar time events and calculates when they elapse next. Use it to verify your OnCalendar expression before deploying the timer.

$ systemd-analyze calendar 'Wed *-*-* 14:00' --iterations=3
  Original form: Wed *-*-* 14:00
Normalized form: Wed *-*-* 14:00:00
    Next elapse: Wed 2025-05-21 14:00:00 UTC
       From now: 1 day left
   Iteration #2: Wed 2025-05-28 14:00:00 UTC
       From now: 8 days left
   Iteration #3: Wed 2025-06-04 14:00:00 UTC
       From now: 15 days left

See here for more on time expressions.

2.4.4. Timer management

After copying the unit file:

• List unit files to verify that systemd sees the new file (list-unit-files)
• Inspect the unit file to verify its content (cat)
• Reload unit files so that systemd picks up any changes (daemon-reload)
• Start the timer (start)
• Check the status to confirm it is active and waiting (status)
• List active timers to see the next trigger time (list-timers)
• Enable the timer so it starts on boot (enable)
• Reboot the system
• Check the status to confirm it survived the reboot (status)

3. Solution

This exercise was proposed on September 8, 2025.

Licenses

Content	License
Code	MIT License
Text	Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International