Because InnoDB is a multi-versioned database, it must keep information about old versions of rows in the tablespace. This information is stored in a data structure called a rollback segment after an analogous data structure in Oracle.

Internally, InnoDB adds two fields to each row stored in the database. A 6-byte field indicates the transaction identifier for the last transaction that inserted or updated the row. Also, a deletion is treated internally as an update where a special bit in the row is set to mark it as deleted. Each row also contains a 7-byte field called the roll pointer. The roll pointer points to an undo log record written to the rollback segment. If the row was updated, the undo log record contains the information necessary to rebuild the content of the row before it was updated.

InnoDB uses the information in the rollback segment to perform the undo operations needed in a transaction rollback. It also uses the information to build earlier versions of a row for a consistent read.

Undo logs in the rollback segment are divided into insert and update undo logs. Insert undo logs are needed only in transaction rollback and can be discarded as soon as the transaction commits. Update undo logs are used also in consistent reads, and they can be discarded only after there is no transaction present for which InnoDB has assigned a snapshot that in a consistent read could need the information in the update undo log to build an earlier version of a database row.

You must remember to commit your transactions regularly, including those transactions that only issue consistent reads. Otherwise, InnoDB cannot discard data from the update undo logs, and the rollback segment may grow too big, filling up your tablespace.

The physical size of an undo log record in the rollback segment is typically smaller than the corresponding inserted or updated row. You can use this information to calculate the space need for your rollback segment.

In the InnoDB multi-versioning scheme, a row is not physically removed from the database immediately when you delete it with an SQL statement. Only when InnoDB can discard the update undo log record written for the deletion can it also physically remove the corresponding row and its index records from the database. This removal operation is called a purge, and it is quite fast, usually taking the same order of time as the SQL statement that did the deletion.

In a scenario where the user inserts and deletes rows in smallish batches at about the same rate in the table, it is possible that the purge thread will start to lag behind, and the table grows bigger and bigger, making everything disk-bound and very slow. Even if the table would carry just 10 MB of useful data, it may grow to occupy 10 GB with all the dead rows. In such a case, it would be good to throttle new row operations, and allocate more resources for the purge thread.

The InnoDB transaction system maintains a list of transactions that have delete-marked index records by UPDATE or DELETE operations. Let the length of this list be purge_lag.

Starting with MySQL/InnoDB-4.1.6, there is a startup option and settable global variable innodb_max_purge_lag, which is zero by default. When this parameter is nonzero, InnoDB may delay new row operations. When the purge_lag exceeds innodb_max_purge_lag, each INSERT, UPDATE and DELETE operation will be delayed by purge_lag/innodb_max_purge_lag*10-5 milliseconds. The delay is computed in the beginning of a purge batch, every ten seconds. The operations will not be delayed if purge cannot run because of an old consistent read view that could see the rows to be purged. A typical setting for a problematic workload might be 1 million, assuming that our transactions are small, only 100 bytes in size, and we can allow 100 MB of unpurged rows in our tables.