Package com.atlassian.confluence.upgrade.upgradetask

Class CheckAndFixSQLServerCatalogNameInJDBCConnectionUrlUpgradeTask

java.lang.Object

com.atlassian.confluence.upgrade.AbstractUpgradeTask

com.atlassian.confluence.upgrade.upgradetask.CheckAndFixSQLServerCatalogNameInJDBCConnectionUrlUpgradeTask

All Implemented Interfaces:

BackupSupport, UpgradeTask, UpgradeTaskInfo, org.springframework.beans.factory.Aware, org.springframework.beans.factory.BeanNameAware

public class CheckAndFixSQLServerCatalogNameInJDBCConnectionUrlUpgradeTask
extends AbstractUpgradeTask

This task will check, and fix the catalog name being used in the JDBC connection url used in confluence.cfg.xml to be correctly cased. Even though the JDBC connection url is not case sensitive, we must correct the case to be the actual case used by the database, as hibernate 5 uses the catalog name in a case sensitive query to obtain metadata, and an incorrectly cased catalog name causes an sql exception: java.sql.SQLException: The database name component of the object qualifier must be the name of the current database.

Since:

6.2.2

Constructor Summary

Constructors

Constructor	Description
CheckAndFixSQLServerCatalogNameInJDBCConnectionUrlUpgradeTask(BootstrapManager bootstrapManager, String configFilename, SingleConnectionProvider databaseHelper, com.atlassian.config.ApplicationConfig applicationConfig, TomcatConfigHelper tomcatConfigHelper)

Method Summary

Modifier and Type	Method	Description
boolean	breaksBackwardCompatibility()	Returns true if a new export can't be imported in an older instance.
void	doUpgrade()	This method actually doesn't do anything, the actual work is done inside the validate() method.
String	getBuildNumber()
boolean	runOnSpaceImport()	Returns true if an older Space can't be imported in a new instance without running this task.
void	validate()	We perform the actual task of changing the jdbc connection url here, even though this is meant to only validate, because the schema check and validation will fail when the catalog name is wrongly cased.

Methods inherited from class com.atlassian.confluence.upgrade.AbstractUpgradeTask

addAllErrors, addError, addError, getConstraint, getErrors, getName, getShortDescription, isDatabaseUpgrade, setBeanName, setBuildNumber

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

CheckAndFixSQLServerCatalogNameInJDBCConnectionUrlUpgradeTask

public CheckAndFixSQLServerCatalogNameInJDBCConnectionUrlUpgradeTask(BootstrapManager bootstrapManager,
                                                                     String configFilename,
                                                                     SingleConnectionProvider databaseHelper,
                                                                     com.atlassian.config.ApplicationConfig applicationConfig,
                                                                     TomcatConfigHelper tomcatConfigHelper)

Method Detail

getBuildNumber

public String getBuildNumber()

Specified by:

getBuildNumber in interface UpgradeTaskInfo

Overrides:

getBuildNumber in class AbstractUpgradeTask

Returns:

The build number that this upgrade is applicable to

runOnSpaceImport

public boolean runOnSpaceImport()

Description copied from interface: BackupSupport

Returns true if an older Space can't be imported in a new instance without running this task. For example:

• A task updating macro names in the BodyContent table would be blocking.
• Tasks which updates data related to the space would be blocking.
• A task upgrading the user table wouldn't be blocking.
• Adding a mandatory column on space-related content breaks space import

Note that tasks don't run on space import yet, so we just reject the import in this case.

breaksBackwardCompatibility

public boolean breaksBackwardCompatibility()

Description copied from interface: BackupSupport

Returns true if a new export can't be imported in an older instance.

Breaking compatibility means a snapshot of the new version will not work at all with the previous version. For example:

• A destructive operation (Some data is replaced by another) breaks backwards compatibility
• Adding a optional column does NOT break backwards compatibility
• Data is copied to another column doesn't breaks backwards compatibility
• A build number incrementation doesn't breaks backwards compatibility

The best way to test is whether a newer export can be imported (with fully working features) in an older instance.

doUpgrade

public void doUpgrade()

This method actually doesn't do anything, the actual work is done inside the validate() method.

validate

public void validate()
              throws Exception

We perform the actual task of changing the jdbc connection url here, even though this is meant to only validate, because the schema check and validation will fail when the catalog name is wrongly cased. The doUpgrade() method is run too late in the schema upgrade process (as the failure occurs as soon as ConfluenceSchemaHelper.validateSchemaUpdateIfNeeded() is run).

Specified by:

validate in interface UpgradeTask

Overrides:

validate in class AbstractUpgradeTask

Throws:

Exception - if validation fails