Blog

In the previous part of our Budget Website Series we have seen how to get you on the internet and what are the prerequisites of the same. This part of the series assumes that you have already registered for your domain name and signed up in Digital Ocean. In this part we will see how to create a new Droplet (VPS) on Digital Ocean and how to connect your domain name. So, lets get started.

Creating a Droplet

Droplet is a computing component in Digital Ocean. This is nothing but the actual cloud server. Droplets are resizable, which means that you have the flexibility to resize server parameters such as Hard Disk Space, Memory and so on without rebuilding the whole server. Of-course your billing will change based on what configuration you are selecting.

To create a new Droplet login to your Digital Ocean account and click on Create Droplet button on the header.

In the last 2 parts of our Budget Website series we have worked on getting a domain name and a hosting server and how to point your domain to the hosting server. In this part we will look at how to actually power up your hosting server so that it can start serving your website.

We will be using WordPress for our content management system to create our budget website. WordPress is a free and open source platform to create websites and blogs and is very easy to use. There is a lot of material on the internet on using, customizing and managing WordPress sites. There are a number of various themes and plugins both free and premium which can be used on WordPress for your specific needs.

The secret sauce to run a decent performance WordPress website on an average server hardware is to use the Nginx web server. Nginx is again a free and open source web server and is pretty light weight. Although configuring Nginx web server can be tricky, but we are in luck as there is a GUI based control panel — Vesta which will do most of the heavy lifting for us.

The main requirements for WordPress are PHP and MySQL. Vesta Control Panel takes care of installing all the required components plus some additional features to harden and secure your server. Let’s start with installing Vesta on our hosting server.

Once you are in the /tmp directory download Vesta package using the following command. This command will always download the latest version of Vesta

$ curl -O http://vestacp.com/pub/vst-install.sh

After the package is downloaded, while still in the /tmp folder, use the below command to start Vesta Control Panel installation.

$ bash vst-install.sh -f

Alright we are now in the last part of our budget website tutorial series. In the last part we have seen how to install and configure all the requirements for the website on your Digital Ocean Droplet. Now in this part we will see how to create a database and install WordPress on your brand new Droplet.

We need to do some minor configuration for WordPress. Once domain is created, click on edit on the domain list page. On the edit domain page scroll down to “Web Template” and select “wordpress2″ from the drop-down.

[BONUS] You can optionally check the “SSL Support” and select “Lets Encrypt Support”. This will add an SSL certificate with https support to your budget website with Green Secure lock. The SSL certificate is issued by Let’s Encrypt and its FREE!

Once done hit save and your web folder and configuration is ready.

Create Database

Now lets create the database. WordPress requires a MySQL database to store all the content, data, settings etc. Creating a database is similar to creating a web domain. Click on the “DB” option on the top and then click on the Green plus site to create a new database.

On the Add Database page enter the database name and the username and password for the database user. You can leave all the other options as it is. Note, Vesta will automatically add the prefix admin_ to your database name and database user. Vesta installs phpMyAdmin as well by default. You can access phpMyAdmin from http://yourdomain.com/phpmyadmin. Note down all the details that you have used to create the database as we will need this to install WordPress for your budget website.

Now go to your browser and visit your domain. You will be presented with the WordPress installation screen. The first step will be to select your language. Select your preferred language and click on Continue.

After WordPress has finished its installation it will ask you for your budget website details and admin user details. These admin details will be used to access your WordPress Dashboard.

As a security precaution never use common words such as admin or administrator for the WordPress admin username. Also, make sure to use a strong and secured password.

That’s All Folks! You have now successfully setup your budget website now. You can login to your WordPress Dashboard on http://yourdomain.com/wp-admin — where yourdomain.com is the domain name that you have installed WordPress on.

You can now go ahead and install some theme, plugins and create pages for your website now.

Stay tuned for some more advanced articles for WordPress, DevOps and much more.

Level: Intermediate

Technologies: VestaCP, WordPress

post via Codincafe

HMTL Data Tables in emails are important for notification’s and other places to display information cleanly. Since only in-line CSS is rendered properly across all email clients its imperative that HTML data table has a proper markup to render it cleanly. So, here is a small code snippet for an HTML data table that renders beautifully in emails.

<table width="99%" border="0" cellpadding="1" cellspacing="0" bgcolor="#EAEAEA">
  <tbody>
    <tr>
      <td>
        <table width="100%" border="0" cellpadding="5" cellspacing="0" bgcolor="#FFFFFF">
          <tbody>
            <tr bgcolor="#EAF2FA">
              <td colspan="2">
                <font style="font-family:sans-serif;font-size:12px"><strong>Name</strong></font>
              </td>
            </tr>
            <tr bgcolor="#FFFFFF">
              <td width="20">&nbsp;</td>
              <td>
                <font style="font-family:sans-serif;font-size:12px">Lead Name</font>
              </td>
            </tr>
            <tr bgcolor="#EAF2FA">
              <td colspan="2">
                <font style="font-family:sans-serif;font-size:12px"><strong>Email</strong></font>
              </td>
            </tr>
            <tr bgcolor="#FFFFFF">
              <td width="20">&nbsp;</td>
              <td>
                <font style="font-family:sans-serif;font-size:12px"><a href="mailto:email@example.com" target="_blank">email@example.com</a></font>
              </td>
            </tr>
            <tr bgcolor="#EAF2FA">
              <td colspan="2">
                <font style="font-family:sans-serif;font-size:12px"><strong>Phone</strong></font>
              </td>
            </tr>
            <tr bgcolor="#FFFFFF">
              <td width="20">&nbsp;</td>
              <td>
                <font style="font-family:sans-serif;font-size:12px">1234567890</font>
              </td>
            </tr>
            <tr bgcolor="#EAF2FA">
              <td colspan="2">
                <font style="font-family:sans-serif;font-size:12px"><strong>Message</strong></font>
              </td>
            </tr>
            <tr bgcolor="#FFFFFF">
              <td width="20">&nbsp;</td>
              <td>
                <font style="font-family:sans-serif;font-size:12px">This is a message</font>
              </td>
            </tr>
          </tbody>
        </table>
      </td>
    </tr>
  </tbody>
</table>

The above code snippet generates an HTML data table email something like this

MySQL has had the unsaid title of “Mother of all Database”, since more than a decade now. It’s a great tool for all size of projects from small to enterprise level applications. Even after all the popularity and wide usage there are some tasks which are better left for more efficient tools. Indexing data in MySQL is rather weak due to a whole lot of reasons. Also, MySQL get bloated pretty quickly which slows it down further.

One of the ways to index MySQL data properly for improved performance on search and analytics is to store the data to be index in a NoSQL type of database. These databases are not relational-type and they store data in a flat format like a tuple.

A perfect tool to help achieve efficient indexing of data is — Elasticsearch. It does an impeccable job of what it does which makes searching and analytics operations much more faster. In one of our previous article we have seen how to install configure a basic setup of Elasticsearch on Ubuntu. In this article we will see how you can actually stream data from MySQL to Elasticsearch with Logstash for indexing which can further be used for search and analytics.

Logstash was initially created for log collection, processing, storage and searching activities. These logs were originally system logs, but developers quickly found out that it can also be used to structure the data in a way that Elasticsearch can understand and store.

Now, it looks like we have all the answers to index MySQL data in Elasticsearch with Logstash right? Wrong. There is a teeny tiny problem to connect MySQL with Logstash. Logstash is originally developed on Java and MySQL requires a JDBC driver known as Connector/J separately so that Logstash can fetch data from it.

Although, this seems a lot of work, its actually quite simple. So, now since we have all the ingredients lets start cooking.

This article assumes that you already have an Elasticsearch instance running on port 9200 and MySQL instance running on port 3306. This article is written for a Ubuntu server

The MySQL Story

Lets take a look at a simple MySQL database

CREATE TABLE IF NOT EXISTS `student` (
 `id` INT NOT NULL AUTO_INCREMENT,
 `first_name` VARCHAR(255) NOT NULL,
 `last_name` VARCHAR(255) NOT NULL,
 PRIMARY KEY (`id`));

CREATE TABLE IF NOT EXISTS `marks` (
 `id` INT NOT NULL AUTO_INCREMENT,
 `marks` VARCHAR(45) NOT NULL,
 `student_id` INT NOT NULL,
 PRIMARY KEY (`id`),
 INDEX `fk_marks_student_idx` (`student_id` ASC),
 CONSTRAINT `fk_marks_student`
 FOREIGN KEY (`student_id`)
 REFERENCES `student` (`id`)
 ON DELETE NO ACTION
 ON UPDATE NO ACTION);

Our sample database has 2 tables — students, marks. The marks tables has an indexed Foreign Key student_id that references from the student table. As the data grows in this table MySQL performance decays. To get the marks of a student we have to use the expensive JOIN clause on 2 tables. Hence, it would be much more easier if only the required data is indexed in form of tuples.

MySQL Connector/J

Let’s start with MySQL’ dependencies. To download and extract the MySQL Connector/J package use the following command

$ wget https://dev.mysql.com/get/Downloads/Connector-J/mysql-connector-java-5.1.42.zip && unzip mysql-connector-java-5.1.42.zip

The extracted package directory will consist of a .jar file — mysql-connector-java-5.1.42-bin.jar

Note the path of the connector file as we will be using this path in the logstash configuration file

Logstash to the rescue

Logstash is a standalone application and does not need to be installed. So, it can be downloaded anywhere on a server where Elasticsearch is running. In this example we are going to download Logstash 5.4.1. You can use the following command to download logstash and extract the package

$ wget https://artifacts.elastic.co/downloads/logstash/logstash-5.4.1.zip && unzip logstash-5.4.1.zip

Your logstash package is now extracted in logstash-5.4.1 directory.

Let’s now create a file student_marks.conf with the below configuration.

input {
  jdbc {
    jdbc_connection_string => "jdbc:mysql://localhost:3306/db_name"
    jdbc_user => "db_user"
    jdbc_password => "db_password"
    jdbc_validate_connection => true
    jdbc_driver_library => "/path/to/mysql-connector-java/mysql-connector-java-5.1.42-bin.jar"
    jdbc_driver_class => "com.mysql.jdbc.Driver"
    jdbc_fetch_size => 100
    statement => "SELECT a.id, b.marks FROM student AS a JOIN marks AS b ON b.student_id = a.id"
  }
}
output {
  elasticsearch {
    index => "student_marks"
    document_type => "student_mark"
    document_id => "%{id}"
    hosts => "localhost:9200"
  }
}

In the configuration file there are 2 blocks viz., input and output. In the input block we have specified that the input is going to come from a jdbc connector. The jdbc block contains the MySQL connection, database user and password, and the path to Connector/J which we have downloaded. Another important definition in the jdbc block is statement.  Statement will have the SQL query of the result that you want to index. In the above example the result of the statement is the primary key from the student table and its corresponding marks.

The output block has another block — elasticsearch. This block defines the parameters of the data index. The document_id should be the primary ID column of the indexing data which in our case is the id column form the students table.

Stream data to Elasticsearch with Logstash

The main logstash application file is in /bin within the logstash directory. To start streaming data from MySQL to Elasticsearch use the following command

$ ./logstash -f /path/to/students_marks.conf

The above command will start pushing all the result from MySQL statement to Elasticsearch. Once the streaming is complete use the following command to check the index status. This command shows the total count of data indexed.

$ curl http://localhost:9200/student_marks/_count

The output of the above command will be something like following in JSON

{"count":1172,"_shards":{"total":5,"successful":5,"failed":0}}

In the output you can see that the value of count property shows that the index student_marks has 1172 documents indexed.

That’s it! You can now use the indexed data from Elasticsearch to put your data search or analytics on steroids.

Virtualization softwares have been in the market since ages. But there has always been a race to create more efficient and high performance platforms among the developers. Few years ago, this lead in the development of yet another virtualization platform — Docker.

Docker is a wee bit different than the traditional virtualization platforms. Docker is a “Software Container” platform. This means that unlike VM’s Docker does not bundle the entire Operating System; rather than that it only bundles libraries and settings in its containers.

We had seen installing VestaCP in our budget website tutorial series. VestaCP provides an easy to use web interface to manage and maintain your datacenter. VestaCP comes with PHP-FPM and there are a lot of different ways to efficiently configure it for optimal performance. On one of our datacenter which we manage was consuming more than 80% of memory even with SWAP (Find out more on SWAP space calculation here). The average CPU utilization was also well above 10% for most of the time. Hence we started to look at the logs and configurations as to what can we do to optimize the server. We looked at a few options where we stumbled upon an article by Steven Sullivanon reducing PHP-FPM memory usage. The article mentioned is an year old and there are some enhancements in VestaCP which can automate a few process.

Let’s dive into the configuration now. VestaCP has a bunch of templates that are used when you create a new site. One of the configuration template is for PHP-FPM settings. There are 2 template files which you can find in the following locations

Let’s consider a long form with 20 or 30 fields. There are times when you want to consume the data of such form directly as name => valuepairs as an object to either send it via Ajax or manipulate data in jQuery and so on. There is a neat little trick to achieve the same. Below is a small snippet which gets all the form values and converts it into a JavaScript Object using jQuery

Customer Retention is broadly defined as an assessment of the product or service quality provided by a business that measures how loyal its customers are. Customer retention statistics are typically expressed as a percentage of long term clients, and they are important to a business since satisfied retained customers tend to spend more, cost less and make valuable references to new potential customers.

Case Study

Lets take an e-commerce website which wants to check their customer retention over time to make informed decisions. Lets assume that the website maintains all data in MySQL. It has 2 tables as follows: