Use our ready-made MySQL service

Good news!, if you would like persistent data in your wasmedge application, simply follow the usage examples below.

Usage

The wasmedge-mysql API will accept valid secure HTTP requests. Whilst the primary use case is to allow Javascript developers to have persistence in their WasmEdge apps, there is nothing stopping you from storing and loading data (even for testing) using any software which is capable of making secure HTTP requests (i.e. Python, Curl, Ruby, PHP). Below, we only provide explicit code examples of using Curl and Javascript.

Endpoints

/api/store

Store raw bytes and receive a UUID to access that data in the future

Curl example

curl --location --request POST 'https://rpc.ssvm.secondstate.io:8888/api/store \
--header 'Content-Type: application/octet-stream' \
--data-binary '@/home/my_user/my_image_file_on_disk'

Javascript example

var data = "<file contents here>";

var xhr = new XMLHttpRequest();
xhr.withCredentials = true;
xhr.addEventListener("readystatechange", function() {
  if(this.readyState === 4) {
    console.log(this.responseText);
  }
});
xhr.open("POST", "https://rpc.ssvm.secondstate.io:8888/api/store");
xhr.setRequestHeader("Content-Type", "application/octet-stream");
xhr.send(data);

---

As mentioned above, this MySQL service is already set up and ready for your requests. If you would like to set up your own service (on your own hardware) please proceed to the next section called "Runing your own wasmedge-mysql".

---

Run your own wasmedge-mysql - optional

System requirements

Ubuntu 20.04 and above

Update system

sudo apt-get update

Install MySQL

Install MySQL using package manager

sudo apt-get install -y mysql-server

Just FYI: In our case we are using an existing MySQL database installation. From here we will be creating a second_server_data directory and starting the existing MySQL software using different port and so forth.

Create new data directory

sudo mkdir -p /opt/second_server_data/

Change permissions and ownership

sudo chmod --reference /var/lib/mysql /opt/second_server_data/
sudo chown --reference /var/lib/mysql /opt/second_server_data/

Edit apparmor configuration

sudo vi /etc/apparmor.d/usr.sbin.mysqld

Add the following block anywhere between the two parenthesis

# Allow second server
/opt/second_server_data/ r,
/opt/second_server_data/** rwk,
/var/run/mysqld/second_server.pid rw,
/var/run/mysqld/second_server.sock rw,
/var/run/mysqld/second_server.sock.lock rw,
/run/mysqld/second_server.pid rw,
/run/mysqld/second_server.sock rw,
/run/mysqld/second_server.sock.lock rw,

Restart apparmor

sudo service apparmor restart

Initialize new MySQL instance

sudo mysqld --initialize --user=mysql --datadir=/opt/second_server_data

To find the new root password simply tail or vi the /var/log/mysql/error.log file

tail /var/log/mysql/error.log

You will notice a line such as the one below (which tells you the new password)

A temporary password is generated for root@localhost: abcdnewpassword

Secure the MySQL instance

sudo mysql_secure_installation --basedir=/opt/second_server_data

Start new MySQL instance

sudo mysqld_safe --no-defaults --datadir=/opt/second_server_data --port=3315 --mysqlx=0 --socket=/var/run/mysqld/second_server.sock &

Logging in, can be done via socket or via TCP

mysql --socket=/var/run/mysqld/second_server.sock -u root -p

mysql -h 127.0.0.1 -P 3315 -u root -p

Create new root password

ALTER USER 'root'@'localhost' IDENTIFIED WITH mysql_native_password BY 'your_password_here';
FLUSH PRIVILEGES;

Create new database

CREATE DATABASE wasmedge_db;
use wasmedge_db;

Create new table

CREATE TABLE wasmedge_data (
    wasmedge_int INT(6) NOT NULL AUTO_INCREMENT,
    wasmedge_id BINARY(16) NOT NULL,
    wasmedge_blob LONGBLOB NOT NULL,
    PRIMARY KEY (wasmedge_int)
);

Create new local user

CREATE USER 'wasmedge_user'@'localhost' IDENTIFIED BY 'your_password_here';
GRANT ALL PRIVILEGES ON wasmedge_db . * TO 'wasmedge_user'@'localhost';
FLUSH PRIVILEGES;
ALTER USER 'wasmedge_user'@'localhost' IDENTIFIED WITH mysql_native_password BY 'your_password_here';
FLUSH PRIVILEGES;

Fetch

git clone https://github.com/second-state/wasmedge-mysql.git
cd wasmedge-mysql

Create config (.env) file

Open a new file called .env.

vi `.env`

Place the following configuration in that file and save.

server_name=rpc.ssvm.secondstate.io
host=0.0.0.0
port=8888

db_host=localhost
db_port=3315
db_user=wasmedge_user
db_password=your_password_here
db_name=wasmedge_db

Automatically install the npm packages that ship with this build's package.json file.

npm install

Start

node index.js

Start using forever

forever start index.js

---

Design discussion - number 3 currently in play

1) Direct remote access - not recommended!

Accessing the database remotely will always require the caller to have a mysql client. Further, it will require that the caller stores the password. Both of these facts make native remote access not practicable in a shared environment. Instead, we are going to implement a simple API which will accept secure HTTP requests and perform MySQL read/write transactions on behalf of the caller.

You could create a new external user like this ...

CREATE USER 'wasmedge_remote_user'@'0.0.0.0' IDENTIFIED BY 'your_password_here';
GRANT ALL PRIVILEGES ON wasmedge_db . * TO 'wasmedge_remote_user'@'0.0.0.0';
FLUSH PRIVILEGES;
ALTER USER 'wasmedge_remote_user'@'0.0.0.0' IDENTIFIED WITH mysql_native_password BY 'your_password_here';
FLUSH PRIVILEGES;

You could test remote access like this ...

Download the MySQL client from dev.mysql.com.

Run the following command on a remote machine

mysql -P 3315 -u wasmedge_remote_user -h 27.33.80.26 -p

However, the above direct remote access solution means that all users are sharing a common username and password.

2) 3rd-party remote access - not ideal

There are many ways to expose MySQL over HTTP using 3rd-party software like aceql. However these also require that you share/give passwords out to end users.

3) In-house key:value store - working on this

One way to store data would be to allow the caller to store anything as a byte array i.e. a blob.

The following convention would need to be followed:

• a user can make a secure HTTP request to store a blob i.e. anything as a byte array
• the user is responsible encoding/decoding their data type to byte array and back again (depending on their application)
• this allows us to store images natively etc (as apposed to only allowing JSON or String)
• the API will always return a valid UUID when data is stored
• the user will load their data using the UUID which the database provided

Here are some examples of the internal SQL operations

Storing a blob

INSERT INTO wasmedge_data(wasmedge_id, wasmedge_blob) VALUES(UUID_TO_BIN(UUID()), the_byte_array);

Returns 994d926e-2a3e-11ec-a87d-08719041559b

Loading a blob

SELECT wasmedge_blob FROM wasmedge_data where BIN_TO_UUID(wasmedge_id) = '994d926e-2a3e-11ec-a87d-08719041559b';

---

Shutdown database

Please use the following command to shutdown this database instance

sudo mysqladmin -h 127.0.0.1 -P 3315 -u root -p shutdown

References

• https://medium.com/@omkarmanjrekar/running-multiple-mysql-instances-on-ubuntu-4af059aad5ae