azcopy_sync/run_azcopy-multi-dir-multi-container.sh

#!/bin/bash

# Enter the specific 'work directory'Ensure we collect logs and other files in the one place.
cd /opt/AZURE || exit 1

# Configurable variables
# Basic variables need for this scrip[t to operate correctly. 
BUSINESS_HOURS_START=10
BUSINESS_HOURS_END=20
AZURE_ACCOUNT=""
AZURE_SAS=""
LOCK_FILE="/tmp/run_azcopy.lock"

# Arguments
# From the comand line. the first is mandatory. There is a check for it if its not provided.
SOURCE_LIST_FILE="$1"
LOGGING="${2:-false}"  # Default to false
BANDWIDTH_CAP="${3:-0}"  # Default is 0 (no cap)

# Report files
# A bunch of temp files used to generate logs and a simple completion report.
TIMESTAMP=$(date +"%Y%m%d_%H%M%S")
LOG_FILE="azcopy_log_$TIMESTAMP.txt"
COMPLETION_REPORT="completion_report_$TIMESTAMP.txt"

# Ensure source list file is provided
# This the check for source files no ne3ed to run if that is not present. 
if [[ -z "$SOURCE_LIST_FILE" || ! -f "$SOURCE_LIST_FILE" ]]; then
    echo "Usage: $0 <directory_list_file> [log=true|false] [bandwidth_mbps]"
    exit 1
fi

# Lock file to prevent multiple instances
# Part of checks to prevent doubling of running proccesses.
if [[ -f "$LOCK_FILE" ]]; then
    PID=$(cat "$LOCK_FILE")
    if kill -0 "$PID" 2>/dev/null; then
        echo "Another instance (PID $PID) is already running. Exiting..."
        exit 1
    else
        echo "Stale lock file found. Removing..."
        rm -f "$LOCK_FILE"
    fi
fi
echo $$ > "$LOCK_FILE"

# Function to check business hours
# This will ensure we do not run at times when the server might need to do other things.
is_business_hours() {
    HOUR=$(date +%H | sed 's/^0*//') # Remove leading zeros causing errors at morning times
    #HOUR=$(printf "%d" "$(date +%H)")  # Convert to decimal safely / this had an issue for some reason not working
    [[ $HOUR -ge $BUSINESS_HOURS_START && $HOUR -lt $BUSINESS_HOURS_END ]]
}

# Stop if running during business hours
# Uses the previous funcion to kill the procces if needed.
if is_business_hours; then
    echo "Business hours detected ($BUSINESS_HOURS_START:00 - $BUSINESS_HOURS_END:00). Exiting..."
    rm -f "$LOCK_FILE"
    exit 1
fi

echo "Starting sync job at $(date)" | tee -a "$LOG_FILE"

# Read the directory list file into an array
# This is used like that becouse of how cron uses a seperate shell enviroment.
mapfile -t SYNC_JOBS < "$SOURCE_LIST_FILE"

# The actual part of the script that does the job.
# Loops through the array and process each entry
# This also does a check to kill the process if it does not complete before the restriction window apply.
# Also checks if a given container exists will stop if it does not.
for LINE in "${SYNC_JOBS[@]}"; do
    IFS=">" read -r SOURCE_DIR DEST_CONTAINER <<< "$LINE"
    SOURCE_DIR=$(echo "$SOURCE_DIR" | xargs)   # Trim spaces
    DEST_CONTAINER=$(echo "$DEST_CONTAINER" | xargs)  # Trim spaces

    if [[ -z "$SOURCE_DIR" || ! -d "$SOURCE_DIR" ]]; then
        echo "ERROR: Invalid directory: $SOURCE_DIR. Exiting." | tee -a "$LOG_FILE"
        rm -f "$LOCK_FILE"
        exit 1
    fi

    if [[ -z "$DEST_CONTAINER" ]]; then
        echo "ERROR: No destination container specified for $SOURCE_DIR. Exiting." | tee -a "$LOG_FILE"
        rm -f "$LOCK_FILE"
        exit 1
    fi

    DEST_URL="$AZURE_ACCOUNT/$DEST_CONTAINER"
    echo "Syncing $SOURCE_DIR to container: $DEST_CONTAINER" | tee -a "$LOG_FILE"

    # Check if the container exists by attempting to write a small test file
    TEST_FILE="$SOURCE_DIR/.azcopy_test_file"
    touch "$TEST_FILE"
    azcopy cp "$TEST_FILE" "$DEST_URL?$AZURE_SAS" > /dev/null 2>&1
    if [[ $? -ne 0 ]]; then
        echo "ERROR: Destination container $DEST_CONTAINER does not exist or is inaccessible. Exiting." | tee -a "$LOG_FILE"
        rm -f "$TEST_FILE"
        rm -f "$LOCK_FILE"
        exit 1
    fi
    rm -f "$TEST_FILE"

    # Run azcopy in the background (one directory at a time)
    if [[ "$LOGGING" == "true" ]]; then
        azcopy sync "$SOURCE_DIR" "$DEST_URL?$AZURE_SAS" --recursive --cap-mbps "$BANDWIDTH_CAP" | tee -a "$LOG_FILE" &
    else
        azcopy sync "$SOURCE_DIR" "$DEST_URL?$AZURE_SAS" --recursive --cap-mbps "$BANDWIDTH_CAP" > /dev/null 2>&1 &
    fi

    AZCOPY_PID=$!

    # Monitor the process every 30 seconds
    while kill -0 $AZCOPY_PID 2>/dev/null; do
        if is_business_hours; then
            echo -e "\nBusiness hours started! Stopping azcopy..." | tee -a "$LOG_FILE"
            kill $AZCOPY_PID
            wait $AZCOPY_PID 2>/dev/null  # Ensure process stops completely
            rm -f "$LOCK_FILE"
            exit 1
        fi
        sleep 30  # Check every 30 seconds
    done

    # Check if sync failed
    if [[ $? -ne 0 ]]; then
        echo "ERROR: Sync failed for $SOURCE_DIR to $DEST_CONTAINER. Stopping script." | tee -a "$LOG_FILE"
        rm -f "$LOCK_FILE"
        exit 1
    fi

done

echo "All directories synced successfully!" | tee -a "$LOG_FILE"

# Generate completion report
echo "Sync Completed: $(date)" > "$COMPLETION_REPORT"
echo "All directories listed in $SOURCE_LIST_FILE have been synced." >> "$COMPLETION_REPORT"

echo "Completion report generated: $COMPLETION_REPORT"
rm -f "$LOCK_FILE"
exit 0