Introduction

This is a breaking of the recovery process into distinct, isolated steps; not all steps are always performed; not all steps are always performed in the same manner. In this doc we will relate to a single table data recovery.

A few notes

• The InnoDB Tools only work for InnoDB/XtraDB tables.
• There is no guarantee. Even with perfect tools, data is sometimes unrecoverable. For example, data that is overwritten cannot be recovered with these tools. There may be file system specific or physical means to recover overwritten data. This is outside the scope of this project.
• Time is crucial. Best chances are when acting the moment things have gone wrong.
• There is manual work to do. Not everything is automatic.
• Recovery depends on you being familiar with your data. As part of the process you may have to choose between two variations of data. The better you know your data, the more chances you'll get good recovery.

Recovery types

There could be several issues from which you may need data recovery:

• A mistaken DROP TABLE t
• A mistaken DELETE FROM t
• A mistaken TRUNCATE TABLE t
• A mistaken UPDATE t SET c='some value meant for single row'; Ooops, forgot WHERE condition
• etc.

Installation steps

• Getting hold of the code: code is available for download from InnoDB Tools project home page.
• Follow these rules.

Recovery steps

Recovery consists of the following steps, not all of which must take place:

• Getting a hold of an accurate table schema.
• Getting hold of tablespace file(s). Do this as soon as possible after the destroying event. The longer this step is delayed, the less chance data can be fully recovered.
• Splitting shared tablespace file.
• Knowing where your data resides withing the tablespace.
• Isolating table's tablespace data.
• Preparing the code for recovery.
• Recovering tablespace data. This is the core of the recovery process. It is in this step that you must apply your knowledge of the data to refine the recovery results. This step is often performed repeatedly until results are refined. The result of this step is a tab-delimited format of your data.
• Importing tab-delimited format back into table. Done with standard LOAD DATA INFILE.

The above steps consist of three main stages:

1. Getting the raw tablespace.
2. Attempting to extract table data.
3. Importing data back to table.

Ongoing example

These docs will use a recovery example, described here.

Let's look at each of these steps in more detail.

Next: sample damage