Mysql – Make a snapshot of a live theSQL database with theISAM & innoDB tables without locking

I've had good success with MySQL replication and nightly tarballs. For smaller db, the mysql database, and schema I use a combination of scripts designed to use mysqlhotcopy and mysqldump.

InnoDB hot backup is a great commercial product but I'm not sure how it handles mixed tables in the same database. pQd's recommendation for XtraBackup may be good to compare against this.

Others like LVM snapshots and I'd say that's definitely something to consider. Ultimately, a combination of solutions would probably be best.

It's also notable this is an old topic. Between the High Performance MySQL book, the MySQL manual, and previous ServerFault questions-- this has been exhausted on a general basis. See:

• Is there any equivalent of mysqlhotcopy for InnoDB?
• Backup MySQL Server
• How do you usually back up your database?

MySQL use of DiskSpace is quite predictable.

The information_schema can quickly give away how much space is used by both storage engines. However, it is far better to configure InnoDB with innodb_file_per_table. That way, you can micromanage the diskspace of individual InnoDB tables. If you do not have innodb_file_per_table, the ibdata1 will grow and NEVER, EVER SHRINK.

I wrote nice articles about Cleaning Up InnoDB once and for all.

• MySQL InnoDB - innodb_file_per_table cons?
• https://stackoverflow.com/questions/3927690/howto-clean-a-mysql-innodb-storage-engine/4056261#4056261

As for MyISAM, you need to periodically run one of the following:

• OPTIMIZE TABLE myisam-tablename;
• ALTER TABLE myisam-tablename ENGINE=MyISAM; ANALYZE TABLE myisam-tablename;

These will compress MyISAM so that there are not unused data and index pages in the MyISAM table components (.MYD and .MYI)

You can manually monitor the disk space usage with this query:

SELECT IFNULL(B.engine,'Total') "Storage Engine", CONCAT(LPAD(REPLACE(FORMAT(B.DSize/POWER(1024,pw),3),',',''),17,' '),' ',SUBSTR(' KMGTP',pw+1,1),'B') "Data Size", CONCAT(LPAD(REPLACE(FORMAT(B.ISize/POWER(1024,pw),3),',',''),17,' '),' ',SUBSTR(' KMGTP',pw+1,1),'B') "Index Size", CONCAT(LPAD(REPLACE(FORMAT(B.TSize/POWER(1024,pw),3),',',''),17,' '),' ',SUBSTR(' KMGTP',pw+1,1),'B') "Table Size" FROM (SELECT engine,SUM(data_length) DSize,SUM(index_length) ISize,SUM(data_length+index_length) TSize FROM information_schema.tables WHERE table_schema NOT IN ('mysql','information_schema','performance_schema') AND engine IS NOT NULL GROUP BY engine WITH ROLLUP) B,(SELECT 3 pw) A ORDER BY TSize;

This will tell how much space is occupied by data and index for engine storage engine. Notice at the end of the query the clause: (SELECT 3 pw). Change the number to generate the report in different units

(SELECT 0 pw) for Bytes
(SELECT 1 pw) for KiloBytes
(SELECT 2 pw) for MegaBytes
(SELECT 3 pw) for GigaBytes
(SELECT 4 pw) for TeraBytes
(SELECT 5 pw) for PetaBytes (Write me if you every get numbers this high)

UPDATE 2012-05-26 22:29 EDT

When it comes to InnoDB, if you use innodb_file_per_table, you may find that there is a difference between the file size of the .ibd file and the sum of data_length + index_length.

For an InnoDB table mydb.mytable, here is the comparison you should make:

• Get the size of file by running ls -l /var/lib/mysql/mydb/mytable.ibd | awk '{print %5}'
• Get the size of the table from the information schemna's point of view: SELECT data_length+index_length FROM information_schema.tables WHERE table_schema='mydb' AND table_name='mytable';
• If the filesize > (data_length+index_length) * 1.1, then you should drag the table like this: ALTER TABLE mydb.mytable ENGINE=InnoDB;

This will make a temp table, copy only real data pages and index pages into the temp table, delete the original, and rename the temp table back to mydb.mytable. Instant table compression with one command. Please plan all table compressions during off hours.