Migrating Sqlite Databases to MySQL

Recently I had to migrate a relatively simple (no FKs or views) Grafana 4 database of about 20 tables and 80,000 rows from sqlite3 to mysql5.6. Below are some notes I made on the 3 methods I tried.

  1. First up was MySQL Workbench on Mac OS. This required first finding and installing a working ODBC driver for sqlite3, which was a hassle. (The Devart one worked.) Then configuring it in the Mac’s ODBC Manager. After clicking a few UI buttons, the import failed with an error message for each table and Workbench crashed, like usual for the past 2 decades. Tried again with another crash. Moving on …
  2. Next up was Navicat Premium Trial Edition. It was able to create a sql file that was syntactically valid with MySQL but had 3 major problems:
    1. it applied the TEXT cast to most of the string and date values
    2. which then meant it thought it had to declare most of the character and date columns as LONGTEXT and LONGBLOB
    3. which meant the index definitions using those columns were invalid because of length. In other words, a mess.
  3. Finally I used a bash script that enumerated the table names using .tables and did a sql dump using .mode insert on each table of individual INSERT statements. That worked fine.

Some other ideas to consider for more complex migrations are:

  • Use the Amazon database migration tools. There’s at least 3 that I’m aware of.
  • Use the professional-grade pgloader to migrate to Postgresql, then use another tool to pivot to MySQL.

The MySQL grants you need for grafana are:

GRANT USAGE ON grafana.* TO 'grafana'@'%' IDENTIFIED BY 'mypw';
GRANT ALL PRIVILEGES ON grafana.* TO 'grafana'@'%';

There is problem with the Grafana initial table creation script: it uses CREATE IF NOT EXISTS for the tables, but not the indexes, so you see this:

CREATE UNIQUE INDEX `UQE_user_login` ON `user` (`login`);

This entry was posted in MySQL, Open Source, Tech. Bookmark the permalink.

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.