release-scripts/backup/database_backup-script.py
Vikas Nale c77eccf8c6 Modify script to support both windows and ubuntu system
Modify script to upload backups on s3
Add documentation
2025-06-05 15:55:06 +05:30

207 lines
8.3 KiB
Python

import os
import datetime
import subprocess
import logging
import sys
import boto3
from botocore.exceptions import ClientError
from dotenv import load_dotenv # Import load_dotenv
# Load environment variables from .env file
load_dotenv()
# --- Configuration ---
# Variables are now loaded from the .env file using os.getenv()
DB_HOST = os.getenv('DB_HOST')
DB_USER = os.getenv('DB_USER')
DB_PASSWORD = os.getenv('DB_PASSWORD')
DB_NAME_PROD = os.getenv('DB_NAME_PROD')
DB_NAME_STAGE = os.getenv('DB_NAME_STAGE')
DB_NAME_GITA = os.getenv('DB_NAME_GITA')
DB_NAME_MEDIAWIKI = os.getenv('DB_NAME_MEDIAWIKI')
DB_NAME_REDMINE = os.getenv('DB_NAME_REDMINE')
# --- AWS S3 Configuration ---
S3_BUCKET_NAME = os.getenv('S3_BUCKET_NAME')
S3_REGION = os.getenv('S3_REGION')
# --- Platform-Specific Paths ---
# Determine OS and set paths accordingly
if sys.platform.startswith('win'):
# Paths for Windows
# You can also load these from .env if you prefer fine-grained control
BACKUP_DIR = os.getenv('WIN_BACKUP_DIR', "C:/gita/database/backup") # Default if not in .env
MYSQLDUMP_PATH = os.getenv('WIN_MYSQLDUMP_PATH', r'C:\Program Files\MySQL\MySQL Server 8.0\bin\mysqldump.exe')
LOG_FILE = os.getenv('WIN_LOG_FILE', r'C:\gita\backup_log.txt')
GIT_EXECUTABLE = "git" # Assuming git is in PATH on Windows
else:
# Paths for Ubuntu/Linux
BACKUP_DIR = os.getenv('LINUX_BACKUP_DIR', "/var/lib/mysql-backups") # Default if not in .env
MYSQLDUMP_PATH = os.getenv('LINUX_MYSQLDUMP_PATH', "/usr/bin/mysqldump")
LOG_FILE = os.getenv('LINUX_LOG_FILE', "/var/log/mysql_backup.log")
GIT_EXECUTABLE = "git" # Assuming git is in PATH on Linux
# --- Logging Setup ---
# Configure logging to write messages to a file and to the console
logging.basicConfig(
level=logging.INFO,
format='%(asctime)s - %(levelname)s - %(message)s',
handlers=[
logging.FileHandler(LOG_FILE),
logging.StreamHandler()
]
)
# --- Helper Functions (unchanged from previous version) ---
def build_path(database_name):
timestamp = datetime.datetime.now().strftime('%Y%m%d%H%M%S')
backup_filename = f"{timestamp}-{database_name}-database-backup.sql"
backup_path = os.path.join(BACKUP_DIR, backup_filename)
return backup_path
def build_mysqldump_command(database_name):
command = [
MYSQLDUMP_PATH,
f"-h{DB_HOST}",
f"-u{DB_USER}",
f"-p{DB_PASSWORD}",
database_name
]
return command
def start_backup(database):
backup_file_path = build_path(database)
command = build_mysqldump_command(database)
try:
logging.info(f"Starting backup for database: {database} to {backup_file_path}")
with open(backup_file_path, "w", encoding="utf-8") as out_file:
subprocess.run(command, stdout=out_file, check=True, text=True)
logging.info(f"Successfully backed up {database}.")
return backup_file_path
except subprocess.CalledProcessError as e:
logging.error(f"MySQL dump failed for {database}: {e}")
logging.error(f"Command: {' '.join(e.cmd)}")
if e.stderr:
logging.error(f"Stderr: {e.stderr}")
return None
except FileNotFoundError:
logging.error(f"Error: mysqldump not found at '{MYSQLDUMP_PATH}'. Please verify the path.")
if sys.platform.startswith('win'):
logging.error(r"On Windows, ensure MySQL is installed and 'mysqldump.exe' path is correct (e.g., C:\Program Files\MySQL\MySQL Server 8.0\bin\mysqldump.exe)")
else:
logging.error("On Ubuntu/Linux, you might need to install 'mysql-client' package: sudo apt install mysql-client")
return None
except Exception as e:
logging.error(f"An unexpected error occurred during backup of {database}: {e}")
return None
def upload_to_git():
original_cwd = os.getcwd()
try:
logging.info(f"Changing directory to {BACKUP_DIR} for Git operations.")
os.chdir(BACKUP_DIR)
logging.info("Staging all changes in Git repository...")
subprocess.run([GIT_EXECUTABLE, "add", "."], check=True, text=True)
commit_message = f"Automated database backup commit on {datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}"
logging.info(f"Committing changes with message: '{commit_message}'")
subprocess.run([GIT_EXECUTABLE, "commit", "-m", commit_message], check=True, text=True)
logging.info("Pushing changes to remote Git repository...")
subprocess.run([GIT_EXECUTABLE, "push"], check=True, text=True)
logging.info("Backup files successfully pushed to Git repository.")
except subprocess.CalledProcessError as e:
logging.error(f"Git upload failed: {e}")
logging.error(f"Command: {' '.join(e.cmd)}")
if e.stderr:
logging.error(f"Stderr: {e.stderr}")
raise
except FileNotFoundError:
logging.error(f"Error: Git executable ('{GIT_EXECUTABLE}') not found. Ensure Git is installed and in your system's PATH.")
raise
except Exception as e:
logging.error(f"An unexpected error occurred during Git upload: {e}")
raise
finally:
os.chdir(original_cwd)
def upload_to_s3(file_paths):
if not file_paths:
logging.info("No backup files to upload to S3.")
return
# Basic validation for S3 configuration
if not S3_BUCKET_NAME or not S3_REGION:
logging.error("S3_BUCKET_NAME or S3_REGION is not set. Cannot upload to S3. Please check your .env file.")
return
try:
logging.info(f"Attempting to connect to AWS S3 bucket: {S3_BUCKET_NAME} in region: {S3_REGION}")
s3_client = boto3.client('s3', region_name=S3_REGION)
for file_path in file_paths:
if not os.path.exists(file_path):
logging.warning(f"File not found, skipping S3 upload: {file_path}")
continue
s3_object_key = os.path.basename(file_path)
try:
logging.info(f"Uploading {s3_object_key} to s3://{S3_BUCKET_NAME}/{s3_object_key}")
s3_client.upload_file(file_path, S3_BUCKET_NAME, s3_object_key)
logging.info(f"Successfully uploaded {s3_object_key} to S3.")
except ClientError as ce:
logging.error(f"Failed to upload {s3_object_key} to S3: {ce}")
if ce.response['Error']['Code'] == 'AccessDenied':
logging.error("S3 upload access denied. Check your AWS credentials and bucket policy.")
elif ce.response['Error']['Code'] == 'NoSuchBucket':
logging.error(f"S3 bucket '{S3_BUCKET_NAME}' does not exist or you don't have access.")
else:
logging.error(f"S3 ClientError details: {ce.response['Error']['Code']} - {ce.response['Error']['Message']}")
except Exception as e:
logging.error(f"An unexpected error occurred during S3 upload of {s3_object_key}: {e}")
logging.info("All S3 uploads attempted.")
except Exception as e:
logging.critical(f"Failed to initialize S3 client or a critical S3 error occurred: {e}")
# --- Main Execution (unchanged from previous version) ---
if __name__ == "__main__":
logging.info("--- Database Backup Process Started ---")
backup_files_created = []
# Basic validation that essential DB connection variables are loaded
if not all([DB_HOST, DB_USER, DB_PASSWORD, DB_NAME_PROD]):
logging.critical("Missing essential database connection variables. Please check your .env file.")
exit(1)
try:
os.makedirs(BACKUP_DIR, exist_ok=True)
logging.info(f"Ensured backup directory exists: {BACKUP_DIR}")
databases_to_backup = [
DB_NAME_PROD,
DB_NAME_STAGE,
DB_NAME_GITA,
DB_NAME_REDMINE,
DB_NAME_MEDIAWIKI
]
for db_name in databases_to_backup:
file_path = start_backup(db_name)
if file_path:
backup_files_created.append(file_path)
logging.info("Starting Git upload process...")
upload_to_git()
logging.info("Starting S3 upload process...")
upload_to_s3(backup_files_created)
logging.info("--- Database Backup Process Completed Successfully ---")
exit(0)
except Exception as ex:
logging.critical(f"--- Database Backup Process Failed: {ex} ---")
exit(1)