Saturday, May 21, 2011

Migrating Data into SQL Server Using SSMA


SSMA provides flexibility for you to migrate your data with one of the two options below: 

1.            Client side data migration engine : This option migrate data through the client machine where SSMA is installed and provides a quick and easy way to migrate smaller database into SQL Server.
2.            Server side data migration engine: this option migrate data directly from the source database to the target database and should be considered when migrating large database
In order to enable server side data migration engine, you will need the following:

1.            Install SSMA Extension Pack
SSMA extension pack installation file comes in the SSMA download but requires separate installation:
 2.            SQL Server Agent
The data migration operation is initiated from SQL Server through a SQL Agent job.

Note that SSMA provides a warning when connecting to target when SQL Server Agent is not running:  "Common Requirement: SQL Server Agent is not running. You must start SQL Server Agent to use Server-side data migration engine". You can ignore this warning if you do not plan to perform data migration or if you decide to use client side data migration engine.
 You can set the data migration option through the project setting:

1.            Navigate to Tools menu and go to Project Settings
2.            Select the migration engine option from the Migration menu


Thursday, May 19, 2011

Migrating from MySQL to SQL Azure Using SSMA


In this blog post, I will describe how to setup your trial SQL Azure account for your migration project, create a “free” database on SQL Azure and walkthrough differences in the process of using SSMA to migrate the tables from the MySQL Sakila sample database to SQL Azure. For a walkthrough of how to migrate a MySQL database to SQL Server, please refer to the post “MySQL to SQL Server Migration:How to use SSMA”. This blog assumes that you have a local version of the MySQL Sakila-DB sample database already installed and that you have  SQL Server Migration Assistant for MySQL v1.0 (SSMA) installed and configured using the instructions provided in the “MySQL to SQL Server Migration: How to use SSMA” blog post.

Getting Started with SQL Azure

If you don’t have a SQL Azure account, you can get a free trial special at http://www.microsoft.com/windowsazure/offers/ through June 30th, 2011. The trial includes a 1GB Web Edition database. Click on the Activate button to get your account up and running. You’ll log in with your Windows Live ID and then complete a four step wizard. Once you are done, the wizard will take you to the Windows Azure Platform portal. If you miss this trail, stay tuned for additional trial offers for SQL Azure.



The next step is to create a new SQL Azure Server by clicking on the “Create a new SQL Azure Server” option in the Getting Started page. You will be prompted for your subscription name that you created in the wizard, the region where your SQL Azure server should be hosted, the Administrator Login information, and the firewall rules. You will need to configure the firewall rules to specify the IPv4 address for the computer where you will be running SSMA.



Click on the Add button to add your firewall rule. Give it a name and then specify a start and end range. The dialog will display your current IPv4 address so that you can enter it in for your start and end range.


Once your are done with the firewall rules, you can then click on your subscription name in the Azure portal to display the fully qualified server name that was just created for you. It will look something like this: x4ju9mtuoi.database.windows.net. You are going to use this server name for your SSMA project.

Using the Azure Portal to Create a Database

SSMA can create a database as part of the project, but for this blog post I’m going to walk through the process of creating the database using the portal and then use the SSMA feature to place the resulting database in a schema within the database. Within the Azure portal, with the subscription selected, you will click in the Create Database command to start the process.


Enter in the name of the database and then keep the other options as the defaults for your free trial account. If you have already paid for a SQL Azure account, you can use the Web edition to go up to 5 GB or switch to the Business edition to up the size limit between 10GB and 50GB. Once created, you will want to click on the Connection Strings – View button to display the connection string information you will use for your SSMA project shown below. The password value shows just a placeholder value.

You are now ready to setup your SSMA project.

Using SSMA to Migrate a MySQL Database to SQL Azure

Once you start SSMA, you will click on the New Project command shown in the image below, enter in the name of the project, and select the SQL Azure option for the Migration To control.



You will then follow the same processes described in the “MySQL to SQL ServerMigration: How to use SSMA” blog post that I will outline below.
  1. Click on the Connect to MySQL toolbar command to complete the MySQL connection information.
    2.  Expand out the Databases node to expand the Sakila database and check the Tables folder as shown below.
 



      3. Click on the Create Report command in the toolbar button. You can ignore the errors. For information about the specific errors with converting the Sakila database, please refer to the blog post “MySQL to SQL Server Migration: Method for Correcting Schema Issues”. Close the report window.
      4. Click on the Connect to SQL Azure button to complete the connection to your target SQL Azure database. Use the first part of the server name and the user name for the administrator as shown below. If you need to change the server name suffix to match your server location, click on the Tools | Project Settings command, click on the General button in the lower left of the dialog and then click on the SQL Azure option in the tree above. From there you can change the suffix value. 


    5. Expand out the Databases node to see the name of the database created in the SQL Azure portal. You will see a Schemas folder under the database name that will be the target for the Sakila database as shown below. 


  6. With the Tables node selected in the MySQL Metadata Explorer, click on the Convert Schema command to create schema named Sakila containing the Sakila tables within the SSMA project  shown below. 

 

    7. Right click on the Sakila schema above and choose the Synchronize with Database command to write the schema changes to your SQL Azure database and then click OK to confirm the synchronization of all objects. This process creates a SQL Server schema object within your database named Sakila and then all the object from your MySQL database go into that schema.
   8. Select the Tables node for the Sakila database in the MySQL Metadata Explorer and then issue the Migrate Data command from the toolbar. Complete the connection dialog to your MySQL database and the connection dialog to the SQL Azure database to start the data transfer. Assuming all goes well, you can dismiss the Data Migration Report as shown below.






At this point, you now have all of the tables and their data loaded into your SQL Azure database contained in a schema named Sakila.

Validating the Results with the SQL Azure Database Manager

To verify the transfer of the data, you can use the Manage command from the Azure Portal and shown below. Just select the database and press the Manage command.



This will launch the SQL Azure Database Manager program in a new browser window with the Server, Database, and Login information prepopulated in the connection dialog. If you have other SQL Azure databases you want to connect to without having to go to the portal, you can always connect via the URL - https://manage-ch1.sql.azure.com/.
Once connected, you can expand the Tables node and select a table like sakila.film to view the structure of the table. You can click on the Data command in the toolbar to view and edit the table’s data as shown below.

 
The SQL Azure Database Manager will also allow you to write an test queries against your database by Database command and then selecting the New Query button in the ribbon.  To learn more about this tool, check out the MSDN topic – Getting Started with The Database Manager for SQL Azure.




Wednesday, May 18, 2011

MySQL to SQL Server Migration: Method for Correcting Schema Issues


In the blog post MySQL to SQL Server Migration: How to Use SSMA, I showed the basic steps of migrating the MySQL Sakila-DB sample database to SQL Server 2008 R2 Express using the SQL Server Migration Assistant for MySQL v1.0 (SSMA). The challenge is that SSMA identified 9 errors with the assessment report associated with the customer, film, film_text, payment, rental, staff, and store tables with an estimated manual conversion time of 9.5 hours. There are three classes of errors that came for the report: (1) SSMA does not convert FULLTEXT indexes; (2) Cascaded foreign key changed to NO ACTION to avoid circular references; (3) Cascaded foreign key changed to NO ACTION to avoid multiple paths. I will walk through each of these scenarios for the successful migration of the Sakila-DB. 

Getting Started with Migrating MySQL’s Sakila Database to SQL Server

If you followed the blog post MySQL to SQL Server Migration: How to UseSSMA, you will have a database already named Sakila on your .\SQLEXPRESS instance. You will want to go into SQL Server Management Studio, connect to the .\SQLEXPRESS instance, open up the Object Explorer if it’s not already open, right click on the Sakila database and issue the Delete command. Use the option in the  Delete Object dialog to Close existing connections and click OK.

Running the SSMA for MySQL

Once you have started SSMA for MySQL, you will create a new project called SakilaAllTables and then follow the steps to Reconnect to MySQL. Inside of the MySQL Metadata Explorer, expand out the Databases node along with the Sakila database. Just check the Tables box and leave the other objects unchecked as shown below.



Right click on the Tables node and select the Create Report command. You will see a report that shows the 9 errors across the three classes of errors as I described in earlier.



SSMA is essentially telling you that it’s going to ignore the errors and that you’ll need to go back and fix them up.

Migrate the Tables and Data

You will next perform the following steps that I described in MySQL to SQL Server Migration: How to Use SSMA post:
  1. Reconnect to SQL Server. You’ll be prompted again to create the Sakila database as part of the connection process.
  2. Convert Schema. You’ll create the model within the SQL Server Metadata Explorer. You can ignore the errors, warnings and informational messages in the Output window for now.
  3. Synchronize with the Database. You’ll right click on the Tables node under the dbo schema and issue the Synchronize with Database command and click ok to create the tables on the SQL Server instance.
  4. Migrate Data. Finally, you will navigate back up to the Tables node in the MySQL Metadata Explorer and issue the Migrate Data command. Follow the connection dialogs and all 16 tables should populate without errors.

Check Out the Results Using SQL Server Management Studio (SSMS)

If you have SQL Server Management Studio still open after deleting the first version of the Sakila database, right click on the Databases node and issue the Refresh command to see the new Sakila database. I’ll now walk through the steps to create a diagram for the Sakila database.
  1. Go ahead and expand out the Sakila database and you will see one folder called Database Diagrams.
  2. Right click on the Database Diagrams folder and issue the Install Diagram Support command. This command creates the stored procedures in the database so that you can see the relationships between the tables. Click Yes for the confirmation message box.
  3. Right click on the Database Diagrams folder and now issue the New Diagram command.
  4. Within the Add Table dialog, select all of the tables and click the Add button.
Once all the tables are loaded into the diagram, click on the Close button and then execute the Database Diagram | Zoom > To Fit command. You should see a diagram that looks like the one below. 




Migrating MySQL Cascaded Foreign Keys to Avoid Circular References: M2SS0036

Within the SSMA assessment report, one of the errors listed was as follows: “M2SS0036: SQL Server Migration Assistant for MySql Error message: ON UPDATE action was changed to NO ACTION to avoid circular references of cascaded foreign keys. (2) Estimated manual conversion time: 2 hr(s)”. In SQL Server, you can end up with some interesting scenarios where cascading foreign keys could cause some interesting problems. Here is what the help file says about cascading foreign keys:
"The series of cascading referential actions triggered by a single DELETE or UPDATE must form a tree that contains no circular references. No table can appear more than one time in the list of all cascading referential actions that result from the DELETE or UPDATE. Also, the tree of cascading referential actions must not have more than one path to any specified table. Any branch of the tree is ended when it encounters a table for which NO ACTION has been specified or is the default."
SSMA works around the problem by pointing out potential problems and creating the foreign keys using the No Action clause.

Don’t Assume That the Database Design is Correct

Before you go about thinking that you need to start writing table triggers to correct the reported error, you need to ask yourself if the use of cascaded updates is the proper design for the application. Let’s take a look at the Staff table first looking at the MySQL create table statement below.
   1:  CREATE TABLE `staff` (
   2:    `staff_id` tinyint(3) unsigned NOT NULL AUTO_INCREMENT,
   3:    `first_name` varchar(45) NOT NULL,
   4:    `last_name` varchar(45) NOT NULL,
   5:    `address_id` smallint(5) unsigned NOT NULL,
   6:    `picture` blob,
   7:    `email` varchar(50) DEFAULT NULL,
   8:    `store_id` tinyint(3) unsigned NOT NULL,
   9:    `active` tinyint(1) NOT NULL DEFAULT '1',
  10:    `username` varchar(16) NOT NULL,
  11:    `password` varchar(40) CHARACTER SET utf8 COLLATE utf8_bin DEFAULT NULL,
  12:    `last_update` timestamp NOT NULL DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP,
  13:    PRIMARY KEY (`staff_id`),
  14:    KEY `idx_fk_store_id` (`store_id`),
  15:    KEY `idx_fk_address_id` (`address_id`),
  16:    CONSTRAINT `fk_staff_address` FOREIGN KEY (`address_id`) REFERENCES `address` (`address_id`) ON UPDATE CASCADE,
  17:    CONSTRAINT `fk_staff_store` FOREIGN KEY (`store_id`) REFERENCES `store` (`store_id`) ON UPDATE CASCADE
  18:  ) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=utf8
Line 17 contains the foreign key constraint that caused the SSMA conversion error. In this case, the desired behavior just looking at the design of the schema is that when the store_id value changes in the store table, make sure the change gets cascaded into the store_id value for Staff table. In addition, there are cascading update foreign keys against store_id for the Customer and Inventory tables. This seems like a natural thing to do, but you need to ask yourself – do I really need to change the key values if the column doesn’t contain meaningful information? In the case of most of the primary keys in the Sakila database, the values are all some form of an integer value. In the world of database design, this is called a Surrogate key. Since the application doesn’t care about the values, there is no reason to change them.
Although the foreign key on line 16 showed no errors with the cascade update rule for address_id changes from the Address table, there is no reason to use the cascade update rule for the same reason.
The recommended course of action would be to turn off the update cascade option for all foreign keys that are made up of surrogate key values, but there is no harm leaving them at this point.

Migrating MySQL Cascaded Foreign Keys to Avoid Multiple Paths: M2SS0037

In the SSMA assessment report, you will find that four tables have ON UPDATE foreign keys that have issues. These tables include: Film, Payment, Rental, and Customer. Let’s see what’s behind this class of errors.
  • Film table. Has one reported error where the foreign key pointing back to the Language table using the language_id column has a problematic ON UPDATE CASCADE clause. SSMA creates the foreign key, but uses the NO ACTION clause. This happened because there is another foreign key pointing back to the Language table using the Film.original_language_id column that uses the ON UPDATE CASCADE clause. Because both use a surrogate key, the errors can be ignored. If you want to change the film$fk_film_language_original foreign key to not use the ON UPDATE CASCADE clause, you can do the following:
    • In SSMS, click on the New Query command on the toolbar.
    • Paste in the following code and then execute the query.
USE sakila
GO
 
ALTER TABLE dbo.film
    DROP CONSTRAINT film$fk_film_language_original
GO
 
ALTER TABLE dbo.film WITH NOCHECK ADD CONSTRAINT
    film$fk_film_language_original FOREIGN KEY
    (
    original_language_id
    ) REFERENCES dbo.language
    (
    language_id
    ) ON UPDATE  NO ACTION 
     ON DELETE  NO ACTION 
GO
  • Payment table. Each of the three foreign keys on the Payment table use surrogate keys, so the errors ON UPDATE CASCASE clause can be safely ignored.
  • Rental table. Each of the three foreign keys on the Rental table use surrogate keys, so the errors ON UPDATE CASCASE clause can be safely ignored.
  • Customer table. Each of the two foreign keys on the Customer table use surrogate keys, so the errors ON UPDATE CASCASE clause can be safely ignored.

Migrating MySQL FULLTEXT Indexes to SQL Server Full-Text Search: M2SS0035

The free SQL Server 2008 R2 Express with the Advanced Services download package provides Full-Text search capabilities similar to MySQL Full-Text search feature. SSMA for MySQL v1.0 doesn’t support the migration of FULLTEXT indexes, but the process is fairly straight forward. You will need to install the Full-Text feature for the SQL Server instance before beginning.

Converting MySQL FULLTEXT Indexes

MySQL can create FULLTEXT indexes on tables that do not have a unique key index. MySQL can also create more than one FULLTEXT index on a table.
The WITH PARSER option can be used only with FULLTEXT indexes. It associates a parser plug-in with the index if full-text indexing and searching operations need special handling. 
Consider the schema for the Film_text table:
CREATE
    TABLE `film_text`
        (
            `film_id` smallint(6) NOT NULL, 
            `title` varchar(255) NOT NULL, 
            `description` text, 
             PRIMARY KEY  (`film_id`) , 
             FULLTEXT  KEY `idx_title_description` (`title`, `description`)  
        )  ENGINE = MyISAM DEFAULT  CHARSET = utf8;
SSMA created the file_text table, but ignored the FULLTEXT key. To create the SQL Server index, you would use the following code:
USE [sakila];
GO
CREATE FULLTEXT CATALOG [Films] AS DEFAULT;
GO
CREATE FULLTEXT INDEX ON dbo.film_text([description]) KEY INDEX pk_film_text_film_id;
GO
You only need to create the catalog once and then use it for other full-text indexes as needed.