Modularizing Shell Scripts for Process Management

Function Decomposition

The script architecture separates core functionalities into distinct functions:

1. fetch_groups - retrieves all group identifiers from configuration
2. fetch_processes - lists process names like 'nginx httpd mysql datanode'
3. retrieve_process_details - provides comprehensive process data including status, PID, CPU, memory, and start time; accepts an optional process name parameter
4. list_group_processes - returns all processes within a specified group

Configuration File Structure

The configuration file process.cfg defines service groups and their associated processes:

[GROUP_LIST]
WEB
DB
HADOOP
YARN

[WEB]
nginx
httpd

[DB]
mysql
postgresql
oracle

[HADOOP]
datanode
namenode
journalnode

[YARN]
resourcemanager
nodemanager

Implementation Details

Retrieving Group Identifiers

fetch_groups() {
    sed -n '/\[GROUP_LIST\]/,/{.*}/p' "$CONFIG_PATH" | egrep -v '(^$|\[.*\])'
}

Fetching Processes With in Groups

fetch_processes() {
    for group in $(fetch_groups); do
        sed -n "/\[$group\]/,/{.*}/p" "$CONFIG_PATH" | egrep -v '(^$|\[.*\])'
    done
}

Process ID Retrieval

get_pid_by_name() {
    if [ $# -ne 1 ]; then
        return 1
    else
        ps -ef | grep "$1" | grep -v grep | grep -v "$0" | awk '{print $2}'
    fi
}

Detailed Process Information

get_process_info() {
    local pid="$1"
    
    if [ "$(ps -ef | awk -v target_pid="$pid" '$2==target_pid{print}' | wc -l)" -eq 1 ]; then
        status="RUNNING"
    else
        status="STOPPED"
    fi
    
    cpu_usage=$(ps aux | awk -v target_pid="$pid" '$2==target_pid{print $3}')
    mem_usage=$(ps aux | awk -v target_pid="$pid" '$2==target_pid{print $4}')
    start_time=$(ps -p "$pid" -o lstart | grep -v STARTED)
}

Group Vaildation and Process Listing

validate_group() {
    for group in $(fetch_groups); do
        if [ "$group" = "$1" ]; then
            return 0
        fi
    done
    return 1
}

list_group_processes() {
    if validate_group "$1"; then
        sed -n "/\[$1\]/,/{.*}/p" "$CONFIG_PATH" | egrep -v '(^$|^#|\[.*\])'
    else
        echo "Group $1 not found in configuration"
    fi
}

Main Execution Logic

The script supports three operational modes:

1. No arguments: displays status of all configured processes
2. -g GROUP_NAME: shows all processes in the specified group
3. PROCESS_NAME: displays detailed status of the specified process

main() {
    # Print header
    printf "%-20s%-20s%-20s%-20s%-20s%-20s%-20s\n" "ProcessName---" "GroupName---" "Status---" "Pid---" "CPU---" "MEMORY---" "StartTime---"
    
    if [ $# -eq 0 ]; then
        # Show all processes
        for proc in $(fetch_processes); do
            group=$(get_group_for_process "$proc")
            is_valid_process "$proc" && format_output "$proc" "$group"
        done
    elif [ "$1" = "-g" ]; then
        # Show specific groups
        shift
        for group in "$@"; do
            is_valid_group "$group" || continue
            for proc in $(list_group_processes "$group"); do
                is_valid_process "$proc" && format_output "$proc" "$group"
            done
        done
    else
        # Show specific processes
        for proc in "$@"; do
            group=$(get_group_for_process "$proc")
            is_valid_process "$proc" && format_output "$proc" "$group"
        done
    fi
}

Output Formatting

format_output() {
    local process="$1"
    local group="$2"
    
    if ps -ef | grep "$process" | grep -v grep | grep -v "$0" > /dev/null; then
        local pids=$(get_pid_by_name "$process")
        for pid in $pids; do
            get_process_info "$pid"
            printf "%-20s%-20s%-20s%-20s%-20s%-20s%-20s\n" "$process" "$group" "$status" "$pid" "$cpu_usage" "$mem_usage" "$start_time"
        done
    else
        printf "%-20s%-20s%-20s%-20s%-20s%-20s%-20s\n" "$process" "$group" "NULL" "Stopped" "NULL" "NULL" "NULL"
    fi
}

The implemantation handles process validation, group membership checks, and structured output formatting for comprehensive system monitoring.