0

🛠️ DB MIGRATION FOR GOLANG SERVICES, WHY IT MATTERS?

DB Migration, why it matters?

Have you ever faced the situations when you deploy new update on production with updated database schemas, but got bugs after that and need to revert things.... that's when migration comes into place.

Database migration serves several key purposes:

  1. Schema Evolution: As applications evolve, their data models change. Migrations allow developers to systematically update the database schema to reflect these changes, ensuring that the database structure matches the application code.
  2. Version Control: Migrations provide a way to version the database schema, allowing teams to track changes over time. This versioning helps in understanding the evolution of the database and aids in collaboration among developers.
  3. Consistency Across Environments: Migrations ensure that the database schema is consistent across different environments (development, testing, production). This reduces the risk of discrepancies that can lead to bugs and integration issues.
  4. Rollback Capability: Many migration tools support rolling back changes, allowing developers to revert to a previous state of the database if a migration causes issues. This enhances stability during the development and deployment process.
  5. Automated Deployment: Migrations can be automated as part of the deployment process, ensuring that the necessary schema changes are applied to the database without manual intervention. This streamlines the release process and reduces human error.

Applying in golang projects

To create a comprehensive, production-grade setup for a Golang service using GORM with MySQL (To get a Quick Setup) that allows for easy migrations, updates, and rollbacks, you need to include migration tooling, handle database connection pooling, and ensure proper struct definitions. Here’s a complete example to guide you through the process:

Project Structure

/golang-service
|-- main.go
|-- database
|   |-- migration.go
|-- models
|   |-- user.go
|-- config
|   |-- config.go
|-- migrations
|   |-- ...
|-- go.mod

1. Database Configuration (config/config.go)

package config

import (
	"fmt"
	"log"
	"os"
	"time"

	"github.com/joho/godotenv"
	"gorm.io/driver/mysql"
	"gorm.io/gorm"
)

var DB *gorm.DB

func ConnectDB() {
	err := godotenv.Load()
	if err != nil {
		log.Fatal("Error loading .env file")
	}

	// charset=utf8mb4: Sets the character set to utf8mb4, which supports all Unicode characters, including emojis.
	// parseTime=True: Tells the driver to automatically parse DATE and DATETIME values into Go's time.Time type.
	// loc=Local: Uses the local timezone of the server for time-related queries and storage.
	dsn := fmt.Sprintf(
		"%s:%s@tcp(%s:%s)/%s?charset=utf8mb4&parseTime=True&loc=Local",
		os.Getenv("DB_USER"),
		os.Getenv("DB_PASS"),
		os.Getenv("DB_HOST"),
		os.Getenv("DB_PORT"),
		os.Getenv("DB_NAME"),
	)

	db, err := gorm.Open(mysql.Open(dsn), &gorm.Config{})
	if err != nil {
		panic("failed to connect database")
	}

	sqlDB, err := db.DB()
	if err != nil {
		panic("failed to configure database connection")
	}

	// Set connection pool settings
	sqlDB.SetMaxIdleConns(10)
	sqlDB.SetMaxOpenConns(100)
	sqlDB.SetConnMaxLifetime(time.Hour)

	// 1.sqlDB.SetMaxIdleConns(10)
	// Sets the maximum number of idle (unused but open) connections in the connection pool.
	// A value of 10 means up to 10 connections can remain idle, ready to be reused.

	// 2. sqlDB.SetMaxOpenConns(100):
	// Sets the maximum number of open (active or idle) connections that can be created to the database.
	// A value of 100 limits the total number of connections, helping to prevent overloading the database.

	// 3. sqlDB.SetConnMaxLifetime(time.Hour):
	// Sets the maximum amount of time a connection can be reused before it’s closed.
	// A value of time.Hour means that each connection will be kept for up to 1 hour, after which it will be discarded and a new connection will be created if needed.

	DB = db
}

2. Database Migration (database/migration.go)

package database

import (
    "golang-service/models"
    "golang-service/migrations"
    "gorm.io/gorm"
)

func Migrate(db *gorm.DB) {
    db.AutoMigrate(&models.User{})
    // Apply additional custom migrations if needed
}

3. Models (models/user.go)

package models

import "gorm.io/gorm"

type User struct {
    gorm.Model
    Name  string `json:"name"`
}

4. Environment Configuration (.env)

DB_USER=root
DB_PASS=yourpassword
DB_HOST=127.0.0.1
DB_PORT=3306
DB_NAME=yourdb

5. Main Entry Point (main.go)

package main

import (
    "golang-service/config"
    "golang-service/database"
    "golang-service/models"
    "github.com/gin-gonic/gin"
    "gorm.io/gorm"
)

func main() {
    config.ConnectDB()
    database.Migrate(config.DB)

    r := gin.Default()
    r.POST("/users", createUser)
    r.GET("/users/:id", getUser)
    r.Run(":8080")
}

func createUser(c *gin.Context) {
    var user models.User
    if err := c.ShouldBindJSON(&user); err != nil {
        c.JSON(400, gin.H{"error": err.Error()})
        return
    }

    if err := config.DB.Create(&user).Error; err != nil {
        c.JSON(500, gin.H{"error": err.Error()})
        return
    }

    c.JSON(201, user)
}

func getUser(c *gin.Context) {
    id := c.Param("id")
    var user models.User

    if err := config.DB.First(&user, id).Error; err != nil {
        c.JSON(404, gin.H{"error": "User not found"})
        return
    }

    c.JSON(200, user)
}

6. Explanation:

  • Database Config: Manages connection pooling for production-grade performance.
  • Migration Files: (in migrations folder) Helps in versioning the database schema.
  • GORM Models: Maps database tables to Go structs.
  • Database Migrations: (in database folder) Custom logic for altering tables over time, allowing for easy rollbacks.
  • Testing: You can create integration tests for this setup using httptest and testify.

7. Create First Migration

  1. For production environments, we could use a migration library like golang-migrate to apply, rollback, or redo migrations.

    Install golang-migrate:

    go install -tags 'mysql' github.com/golang-migrate/migrate/v4/cmd/migrate@latest
    
  2. Generate migrate files for users table

    migrate create -ext=sql -dir=./migrations -seq create_users_table
    

    After running the command we'll get a pair of .up.sql (to update schema) and down.sql (for potential rollback later) . The number 000001 is the auto generated index of migration.

    /golang-service
    |-- migrations
    |   |-- 000001_create_users_table.down.sql
    |   |-- 000001_create_users_table.up.sql
    

    Add relevant sql command to .up file , and .down file.

    000001_create_users_table.up.sql

    CREATE TABLE users (
    id BIGINT AUTO_INCREMENT PRIMARY KEY,
    name VARCHAR(255) NOT NULL,
    created_at DATETIME,
    updated_at DATETIME,
    deleted_at DATETIME);
    

    000001_create_users_table.down.sql

    DROP TABLE IF EXISTS users;
    

    Run the up migration and apply changes to the database with the following command (-verbose flag to see more log details):

    migrate -path ./migrations -database "mysql://user:password@tcp(localhost:3306)/dbname" -verbose up
    

    In case we got issue with migration we can use the following command to see the current migration version and its status:

    migrate -path ./migrations -database "mysql://user:password@tcp(localhost:3306)/dbname" version
    

    If we have a broken migration for some reasons we can consider to use the force (use carefully) command with the version number of the dirty migration. If the version is 1 (could check it in migrations or schema_migrations table), we would run:

    migrate -path ./migrations -database "mysql://user:password@tcp(localhost:3306)/dbname" force 1
    

8. Changing schemes

  1. At some point in time, we might like to add new features and some of those might require chaning data schemes, for instance we'd like to add email field to users table. We'd do it as following.

    Make a new migration for adding email column to users table

    migrate create -ext=sql -dir=./migrations -seq add_email_to_users
    

    Now we have a new pair of .up.sql and .down.sql

    /golang-service
    |-- migrations
    |   |-- 000001_create_users_table.down.sql
    |   |-- 000001_create_users_table.up.sql
    |   |-- 000002_add_email_to_users.down.sql
    |   |-- 000002_add_email_to_users.up.sql
    
  2. Adding following content to *_add_email_to_users.*.sql files

    000002_add_email_to_users.up.sql

    ALTER TABLE `users` ADD COLUMN `email` VARCHAR(255) UNIQUE;
    

    000002_add_email_to_users.down.sql

    ALTER TABLE `users` DROP COLUMN `email`;
    

    Run the up migration command again to make update to the data schemas

    migrate -path ./migrations -database "mysql://user:password@tcp(localhost:3306)/dbname" -verbose up
    

    We'll also need to update the golang users struct (adding Email field) to keep it in sync with the new schemas..

    type User struct {
     gorm.Model
     Name  string `json:"name"`
     Email string json:"email" gorm:"uniqueIndex"
    }
    

9. Rolling Back Migrations:

In case for some reasons we got bugs with new updated schemas, and we need to rollback, this case we'll use the down command:

migrate -path ./migrations -database "mysql://user:password@tcp(localhost:3306)/dbname" down 1

Number 1 indicates that we want to rollback 1 migration.

Here we also need manually to update golang users struct (remove the Email field) to reflect the data schema changes.

   type User struct {
    gorm.Model
    Name  string `json:"name"`
   }

10. Use with Makefile

To simplify the process of migration and rolling back, we can add a Makefile .

/golang-service
|-- ...
|-- Makefile

The content of Makefile as following.

include .env

create_migration:
	migrate create -ext=sql -dir=./migrations -seq create_new_migration

migrate_up:
	migrate -path=./migrations -database "mysql://${DB_USER}:${DB_PASS}@tcp(${DB_HOST}:${DB_PORT})/${DB_NAME}" -verbose up 1

migrate_down:
	migrate -path=./migrations -database "mysql://${DB_USER}:${DB_PASS}@tcp(${DB_HOST}:${DB_PORT})/${DB_NAME}" -verbose down 1

.PHONY:  create_migration migrate_up migrate_down

Now we can simply run make migrate_up or make migrate_down on CLI to do the migration and the rollback.

11.Considerations:

  • Data Loss During Rollback: Rolling back migrations that delete columns or tables may result in data loss, so always backup data before running a rollback.
  • CI/CD Integration: Integrate the migration process into your CI/CD pipeline to automate schema changes during deployment.
  • DB Backups: Schedule regular database backups to prevent data loss in case of migration errors.

About DB backups

Before rolling back a migration or making changes that could potentially affect your database, here are some key points to consider.

  1. Schema Changes: If the migration involved altering the schema (e.g., adding or removing columns, changing data types), rolling back to a previous migration can result in the loss of any data stored in those altered columns or tables.
  2. Data Removal: If the migration includes commands that delete data (like dropping tables or truncating tables), rolling back will execute the corresponding "down" migration, which could permanently remove that data.
  3. Transaction Handling: If your migration tool supports transactions, the rollback might be safer since changes are applied in a transaction. However, if you manually run SQL commands outside of transactions, there is a risk of losing data.
  4. Data Integrity: If you have modified data in a way that depends on the current schema, rolling back could leave your database in an inconsistent state.

So it’s crucial to back up your data. Here’s a brief guide:

  1. Database Dump:
  • Use database-specific tools to create a full backup of your database. For MySQL, you can use:
    mysqldump -u root -p dbname > backup_before_rollback.sql
    
    This creates a file (backup_before_rollback.sql) that contains all the data and schema of the dbname database.
  1. Export Specific Tables:
  • If you only need to back up certain tables, specify them in the mysqldump command:
    mysqldump -u root -p golang_1 users > users_table_backup.sql
    
  1. Verify the Backup:
  • Ensure that the backup file has been created and check its size or open it to ensure it contains the necessary data.
  1. Store Backups Securely:
  • Keep a copy of the backup in a secure location, such as cloud storage or a separate server, to prevent data loss during the rollback process.

Backups on cloud

To back up your MySQL data when using Golang and deploying on AWS EKS, you can follow these steps:

  1. Use mysqldump for Database Backup:
  • Create a mysqldump of your MySQL database using a Kubernetes cron job.
    mysqldump -h <mysql-host> -u <user> -p<password> <database_name> > backup.sql
    
    Store this in a persistent volume or an S3 bucket.
  1. Automate with Kubernetes CronJob:
  • Use a Kubernetes CronJob to automate the mysqldump process.
  • Example YAML configuration:yaml
    apiVersion: batch/v1
    kind: CronJob
    metadata:
    name: mysql-backup
    spec:
       schedule: "0 2 * * *" # Runs every day at 2 AM
       jobTemplate:
         spec:
           template:
             spec:
               containers:
                 - name: mysql-backup
                   image: mysql:5.7
                   args:
                     - /bin/sh
                     - -c
                     - "mysqldump -h <mysql-host> -u <user> -p<password> <database_name> | aws s3 cp - s3://<your-bucket-name>/backup-$(date +\\%F).sql"
                   env:
                     - name: AWS_ACCESS_KEY_ID
                       value: "<your-access-key>"
                     - name: AWS_SECRET_ACCESS_KEY
                       value: "<your-secret-key>"
               restartPolicy: OnFailure
    
  1. Using AWS RDS Automated Backups (if using RDS):
  • If your MySQL database is on AWS RDS, you can leverage RDS automated backups and snapshots.
  • Set a backup retention period and take snapshots manually or automate snapshots using Lambda functions.
  1. Back Up Persistent Volumes (PV) with Velero:
  • Use Velero, a backup tool for Kubernetes, to back up the persistent volume that holds MySQL data.
  • Install Velero on your EKS cluster and configure it to back up to S3.

By using these methods, you can ensure your MySQL data is regularly backed up and securely stored.

If you found this helpful, let me know by leaving a 👍 or a comment!, or if you think this post could help someone, feel free to share it! Thank you very much! 😃


All Rights Reserved

Viblo
Let's register a Viblo Account to get more interesting posts.