使Python通过SSH隧道连接到MySQL

Question

I'm using MySqldb with Python 2.7 to allow Python to make connections to another MySQL server

import MySQLdb
db = MySQLdb.connect(host="sql.domain.com",
     user="dev", 
      passwd="*******", 
      db="appdb")

Instead of connecting normally like this, how can the connection be made through a SSH tunnel using SSH key pairs?

The SSH tunnel should ideally be opened by Python. The SSH tunnel host and the MySQL server are the same machine.

Answer 1

Only this worked for me

import pymysql
import paramiko
import pandas as pd
from paramiko import SSHClient
from sshtunnel import SSHTunnelForwarder
from os.path import expanduser

home = expanduser('~')
mypkey = paramiko.RSAKey.from_private_key_file(home + pkeyfilepath)
# if you want to use ssh password use - ssh_password='your ssh password', bellow

sql_hostname = 'sql_hostname'
sql_username = 'sql_username'
sql_password = 'sql_password'
sql_main_database = 'db_name'
sql_port = 3306
ssh_host = 'ssh_hostname'
ssh_user = 'ssh_username'
ssh_port = 22
sql_ip = '1.1.1.1.1'

with SSHTunnelForwarder(
        (ssh_host, ssh_port),
        ssh_username=ssh_user,
        ssh_pkey=mypkey,
        remote_bind_address=(sql_hostname, sql_port)) as tunnel:
    conn = pymysql.connect(host='127.0.0.1', user=sql_username,
            passwd=sql_password, db=sql_main_database,
            port=tunnel.local_bind_port)
    query = '''SELECT VERSION();'''
    data = pd.read_sql_query(query, conn)
    conn.close()

Answer 2

I'm guessing you'll need port forwarding. I recommend sshtunnel.SSHTunnelForwarder

import mysql.connector
import sshtunnel

with sshtunnel.SSHTunnelForwarder(
        (_host, _ssh_port),
        ssh_username=_username,
        ssh_password=_password,
        remote_bind_address=(_remote_bind_address, _remote_mysql_port),
        local_bind_address=(_local_bind_address, _local_mysql_port)
) as tunnel:
    connection = mysql.connector.connect(
        user=_db_user,
        password=_db_password,
        host=_local_bind_address,
        database=_db_name,
        port=_local_mysql_port)
    ...

Answer 3

from sshtunnel import SSHTunnelForwarder
import pymysql
import pandas as pd

tunnel = SSHTunnelForwarder(('SSH_HOST', 22), ssh_password=SSH_PASS, ssh_username=SSH_UNAME,
     remote_bind_address=(DB_HOST, 3306)) 
tunnel.start()
conn = pymysql.connect(host='127.0.0.1', user=DB_UNAME, passwd=DB_PASS, port=tunnel.local_bind_port)
data = pd.read_sql_query("SHOW DATABASES;", conn)

credits to https://www.reddit.com/r/learnpython/comments/53wph1/connecting_to_a_mysql_database_in_a_python_script/

Answer 4

If your private key file is encrypted, this is what worked for me:

    mypkey = paramiko.RSAKey.from_private_key_file(<<file location>>, password='password')
    sql_hostname = 'sql_hostname'
    sql_username = 'sql_username'
    sql_password = 'sql_password'
    sql_main_database = 'sql_main_database'
    sql_port = 3306
    ssh_host = 'ssh_host'
    ssh_user = 'ssh_user'
    ssh_port = 22


    with SSHTunnelForwarder(
            (ssh_host, ssh_port),
            ssh_username=ssh_user,
            ssh_pkey=mypkey,
            ssh_password='ssh_password',
            remote_bind_address=(sql_hostname, sql_port)) as tunnel:
        conn = pymysql.connect(host='localhost', user=sql_username,
                               passwd=sql_password, db=sql_main_database,
                               port=tunnel.local_bind_port)
        query = '''SELECT VERSION();'''
        data = pd.read_sql_query(query, conn)
        print(data)
        conn.close()

Answer 5

You may only write the path to the private key file: ssh_pkey='/home/userName/.ssh/id_ed25519' (documentation is here: https://sshtunnel.readthedocs.io/en/latest/ ).

If you use mysql.connector from Oracle you must use a construction cnx = mysql.connector.MySQLConnection( ... Important: a construction cnx = mysql.connector.connect( ... does not work via an SSh! It is a bug. (The documentation is here: https://dev.mysql.com/doc/connector-python/en/connector-python-connectargs.html ).

Also, your SQL statement must be ideal. In case of an error on SQL server side, you do not receive an error message from SQL-server.

import sshtunnel
import numpy as np

with sshtunnel.SSHTunnelForwarder(ssh_address_or_host='ssh_host',
                                  ssh_username="ssh_username",
                                  ssh_pkey='/home/userName/.ssh/id_ed25519',
                                  remote_bind_address=('localhost', 3306),
                                  ) as tunnel:
    cnx = mysql.connector.MySQLConnection(user='sql_username',
                                          password='sql_password',
                                          host='127.0.0.1',
                                          database='db_name',
                                          port=tunnel.local_bind_port)
    cursor = cnx.cursor()
    cursor.execute('SELECT * FROM db_name.tableName;')
    arr = np.array(cursor.fetchall())
    cursor.close()
    cnx.close()

Answer 6

Paramiko is the best python module to do ssh tunneling. Check out the code here: https://github.com/paramiko/paramiko/blob/master/demos/forward.py

As said in comments this one works perfect. SSH Tunnel for Python MySQLdb connection

Answer 7

Best practice is to parameterize the connection variables. Here is how I have implemented. Works like charm!

import mysql.connector
import sshtunnel
import pandas as pd
import configparser

config = configparser.ConfigParser()
config.read('c:/work/tmf/data_model/tools/config.ini')

ssh_host = config['db_qa01']['SSH_HOST']
ssh_port = int(config['db_qa01']['SSH_PORT'])
ssh_username = config['db_qa01']['SSH_USER']
ssh_pkey = config['db_qa01']['SSH_PKEY']
sql_host = config['db_qa01']['HOST']
sql_port = int(config['db_qa01']['PORT'])
sql_username = config['db_qa01']['USER']
sql_password = config['db_qa01']['PASSWORD']

with sshtunnel.SSHTunnelForwarder(
        (ssh_host,ssh_port),
        ssh_username=ssh_username,
        ssh_pkey=ssh_pkey,
        remote_bind_address=(sql_host, sql_port)) as tunnel:
    connection = mysql.connector.connect(
        host='127.0.0.1',
        port=tunnel.local_bind_port,
        user=sql_username,
        password=sql_password)
    query = 'select version();'
    data = pd.read_sql_query(query, connection)
    print(data)
    connection.close()

Answer 8

This works for me:

import mysql.connector
import sshtunnel

with sshtunnel.SSHTunnelForwarder(
    ('ip-of-ssh-server', 'port-in-number-format'),
    ssh_username = 'ssh-username',
    ssh_password = 'ssh-password',
    remote_bind_address = ('127.0.0.1', 3306)
) as tunnel:
    connection = mysql.connector.connect(
        user = 'database-username',
        password = 'database-password',
        host = '127.0.0.1',
        port = tunnel.local_bind_port,
        database = 'databasename',
    )
    mycursor = connection.cursor()
    query = "SELECT * FROM datos"
    mycursor.execute(query)

Answer 9

Someone said this in another comment. If you use the python mysql.connector from Oracle then you must use a construction cnx = mysql.connector.MySQLConnection(... .

Important: a construction cnx = mysql.connector.connect(... does not work via an SSH: This bug cost me a whole day trying to understand why connections were being dropped by the remote server:

with sshtunnel.SSHTunnelForwarder(
        (ssh_host,ssh_port),
        ssh_username=ssh_username,
        ssh_pkey=ssh_pkey,
        remote_bind_address=(sql_host, sql_port)) as tunnel:
    connection = mysql.connector.MySQLConnection(
        host='127.0.0.1',
        port=tunnel.local_bind_port,
        user=sql_username,
        password=sql_password)
    query = 'select version();'
    data = pd.read_sql_query(query, connection)
    print(data)
    connection.close()

Answer 10

If you are using python, and all the username, password, host and port are correct then there is just one thing left, that is using the argument (use_pure=True). This argument uses python to parse the details and password. You can see the doc of mysql.connector.connect() arguments.

with sshtunnel.SSHTunnelForwarder(
    (ssh_host,ssh_port),
    ssh_username=ssh_username,
    ssh_pkey=ssh_pkey,
    remote_bind_address=(sql_host, sql_port)) as tunnel:
connection = mysql.connector.MySQLConnection(
    host='127.0.0.1',
    port=tunnel.local_bind_port,
    user=sql_username,
    password=sql_password,
    use_pure='True')
query = 'select version();'
data = pd.read_sql_query(query, connection)
print(data)
connection. Close()