How to Create a Bash Script to Kill All Processes Related to a Service

When managing a Linux system, you may find yourself needing to quickly kill all processes associated with a particular service. This can happen if the service is not behaving properly, crashes, or does not respond to standard shutdown commands. In this guide, we will see how to create a Bash script to simplify this task.

The idea is to automate the search and killing of all processes associated with a specific service, without having to manually identify each one. I will show you step by step how to create a script that does this efficiently.

There are many reasons why you might want to kill all processes of a service. For example:

Freezes or crashes: The service may stop responding, and despite attempts to restart or stop it, the associated processes continue to behave abnormally.
Updates or changes: Sometimes it is necessary to kill all processes of a service to apply configuration changes or software updates.
Performance issues: In some cases, a service may use excessive system resources (e.g. too much CPU or memory), requiring a forced "reset".

Here is the complete code of the script that allows you to kill all processes associated with a service:


#!/bin/bash

# Check if the user has provided the service name
if [ -z "$1" ]; then
  echo "Usage: $0 <service>"
  exit 1
fi

# Save the service name passed as an argument
SERVICE=$1

# Find processes associated with the service
PIDS=$(pgrep -f "$SERVICE")

# Check if there are processes to kill
if [ -z "$PIDS" ]; then
  echo "No processes found for service: $SERVICE"
  exit 0
fi

# Kill all processes found
echo "Found the following processes for $SERVICE: $PIDS"
echo "Killing processes..."
kill -9 $PIDS

# Check if the processes were killed successfully
if [ $? -eq 0 ]; then
  echo "All processes related to $SERVICE were killed."
else
  echo "Error killing $SERVICE processes."
fi

Script explanation:

Input check: The first thing the script does is check if the user has provided the service name. If no argument is given, the script displays a usage message and stops.
Search for processes: The script uses the pgrep -f command to find all processes that match the service name. The -f option allows you to search not only the process name but also the entire command line, which is useful if the service name is part of a longer command.
Checking for found processes: If no matching processes are found, the script informs the user that there are no processes to kill and exits.
Terminating processes: If any processes are found, the script uses the kill -9 command to force immediate termination. The -9 signal corresponds to SIGKILL, which terminates the process without giving it a chance to handle the signal or clean up resources properly. This is useful for blocked processes.
Success verification: The script checks whether the kill command was successful and informs the user of the result.

Important considerations

Forcing process termination: The -9 (SIGKILL) signal forces processes to terminate immediately, without giving them a chance to properly clean up resources or complete any pending operations. Use this option with caution, especially on critical services.
Using SIGTERM: If you prefer to kill processes more "gracefully", you can replace kill -9 with kill, which sends the SIGTERM signal. This gives processes the ability to kill in an orderly manner:
```
kill $PIDS
```
Permissions check: You may need administrator privileges to kill some processes, so be sure to run the script with sudo if necessary:
```
sudo ./kill_service.sh service
```

Conclusion

This simple Bash script allows you to automate the search and killing of all processes associated with a specific service. It is useful in situations where standard service management commands fail or the service is hung. Remember to always use the SIGKILL signal with caution, as it immediately kills processes without allowing an orderly shutdown.