The Master-DB is corrupted
There are two main ways that the data on the Master-DB can become corrupted.
- Data corruption due to the modification of the data stored on the disk (not due to SQL queries). For example, data mutations due to disk malfunctions, filesystem errors, etc.
- Data corruption due to modifications of the database caused by unwanted or incorrect SQL statements. For example, a bug in an application might "destroy/corrupt" the contents of the database.
Verification: How do I know my system is in this situation?
If you think that your database is corrupted, you have probably already noticed odd behavior that's probably related to your database. If it's disk or filesystem corruption you'll probably see error messages in the instance log. If the corruption is caused by a bug in the application, you'll probably notice that data is missing or get application errors complaining about missing relationships, records not found, or failed transactions.
The plan of action changes dramatically depending on the two cases:
Case A - Only my master DB is corrupted
When the data corruption is local to the Master-DB such as disk errors, filesystem errors, etc., it's likely that the problem does not affect the Slave-DB instance. In this case, you can use the Slave-DB for recovery since it has the most recent copy of the non-corrupted data. In this situation, the recovery is the same as when The Master-DB instance failed.
Required actions: What scripts should I run to solve the situation?
Case B - Data on both master and slave DB is corrupted
- © Copyright 2015 RightScale Technical Support