Spring supports programmatic transaction which give us fine grained control over TX management. According to Spring Documentation, One can use programmatic TX management by:
1. utilizing Spring's TransactionTemplate:
transactionTemplate.execute(new TransactionCallbackWithoutResult() {
protected void doInTransactionWithoutResult(TransactionStatus status) {
try {
updateOperation1();
updateOperation2();
} catch (SomeBusinessExeption ex) {
status.setRollbackOnly();
}
} });
2. leveraging PlatformTransactionManager directly(inject a PlatformTransactionManager implementation into DAO):
DefaultTransactionDefinition def = new DefaultTransactionDefinition();
def.setName("SomeTxName");
def.setPropagationBehavior(TransactionDefinition.PROPAGATION_REQUIRED);
//txManager is a reference to PlatformTransactionManager
TransactionStatus status = txManager.getTransaction(def);
try {
updateOperation1();
updateOperation2();
}
catch (MyException ex) {
txManager.rollback(status);
throw ex;
}
txManager.commit(status);
for the sake of simplification, let's say we are dealing with JDBC database operation.
I am wondering for any database operations happened at updateOperation1(),updateOperation2() in the second snippet, either it is implemented with JDBCTemplate or JDBCDaoSupport, if not, the operation is actually not performed within any transaction, is it?
My analysis is that if we don't use JDBCTemplate or JDBCDaoSupport, we inevitably will create/retrieve connection from datasource management. the connection we get is of course not the connection used by PlatformTransactionManager underlying to manage transaction.
I dug Spring source code and skim related class found that PlatformTransactionManager will try to retrieve a connection contained in ConnectionHolder which in return retrieved from TransactionSynchronizationManager. I also found JDBCTemplate and JDBCDaoSupport, also try to get connection with similar routine from TransactionSynchronizationManager.
Because TransactionSynchronizationManager manages many resource including connection per thread(basically use Threadlocal to ensure one thread get its own unique instance of the managed resource)
So I think the connection retrieved by PlatformTransactionManager and JDBCTemplate or JDBCDaoSupport is just same, this can explain how spring programmatic transaction ensure updateOperation1(),updateOperation2() were guarded by transaction.
Is my analysis correct? if it is, why Spring documentation hasn't emphasized this caveat?
Yes, it's correct.
Any code that uses raw Connections should obtain them from the DataSource in special way in order to participate in transactions managed by Spring (12.3.8 DataSourceTransactionManager):
Application code is required to retrieve the JDBC connection through DataSourceUtils.getConnection(DataSource) instead of Java EE's standard DataSource.getConnection.
Another option (if you cannot change code that calls getConnection()) is to wrap your DataSource with TransactionAwareDataSourceProxy.
Related
My application works with multi datasources and 2 databases Oracle and PostgreSQL (I dont need global transaction) .
I dont know which transaction manager to use. Both have some advantages and disadvantages.
Atomikos suppport global transaction which I dont need and log some information about transaction to file system which I want to avoid:
public void setEnableLogging(boolean enableLogging)
Specifies if disk logging should be enabled or not. Defaults to true.
It is useful for JUnit testing, or to profile code without seeing the
transaction manager's activity as a hot spot but this should never be
disabled on production or data integrity cannot be guaranteed.
advantages is that it use just one transaction manager
When using DataSourceTransactionManager I need one per dataSource
#Bean
#Primary
DataSourceTransactionManager transactionManager1() {
DataSourceTransactionManager transactionManager = new DataSourceTransactionManager();
transactionManager.setDataSource(dataSource1());
return transactionManager;
}
#Bean
DataSourceTransactionManager transactionManager2() {
DataSourceTransactionManager transactionManager = new DataSourceTransactionManager();
transactionManager.setDataSource(dataSource2());
return transactionManager;
}
this is problem because I need to specify name of tm in annotation:
#Transactional("transactionManager1")
public void test() {
}
but I dont know it because in runtime I can switch in application which database to use.
is there some other options or I am missing something in this two transaction manager ?
You should solve this as option 2, using one DataSourceTransactionManager per data source. You will need to keep track of the transaction manager for each data source.
One thing additionally, if you need to be able to rollback transactions on both databases, you will have to set up a ChainedTransactionManager for both.
I am having trouble finding information about this issue I am running into. I am interested in implementing row level security on my Postgres db and I am looking for a way to be able to set postgres session variables automatically through some form of an interceptor. Now, I know that with hibernate you are able to do row-level-security using #Filter and #FilterDef, however I would like to additionally set policies on my DB.
A very simple way of doing this would be to execute the SQL statement SET variable=value prior to every query, though I have not been able to find any information on this.
This is being used on a spring-boot application and every request is expected to will have access to a request-specific value of the variable.
Since your application uses spring, you could try accomplishing this in one of a few ways:
Spring AOP
In this approach, you write an advice that you ask spring to apply to specific methods. If your methods use the #Transactional annotation, you could have the advice be applied to those immediately after the transaction has started.
Extended TransactionManager Implementation
Lets assume your transaction is using JpaTransactionManager.
public class SecurityPolicyInjectingJpaTransactionManager extends JpaTransactionManager {
#Autowired
private EntityManager entityManager;
// constructors
#Override
protected void prepareSynchronization(DefaultTransactionStatus status, TransactionDefinition definition) {
super.prepareSynchronization(status, definition);
if (status.isNewTransaction()) {
// Use entityManager to execute your database policy param/values
// I would suggest you also register an after-completion callback synchronization
// This after-completion would clear all the policy param/values
// regardless of whether the transaction succeeded or failed
// since this happens just before it gets returned to the connection pool
}
}
}
Now simply configure your JPA environment to use your custom JpaTransactionManager class.
There are likely others, but these are the two that come to mind that I've explored.
How can I turn off auto commit and give explicit commit using #transactional?
I have multiple operations and I want them to get committed after all select and update are done. Reason being the select is done based on some field as null and this data is used somewhere and then update is done to some records. So before new records come, I have to change the field as to some value and avoid selection of new data i.e select only those record that got updated
You just need to put the #Transaction annotation around all the work you want to be committed:
#Autowired
private Manager1 manager1;
#Autowired
private Manager2 manager2;
#Transactional
public void doStuff() {
manager1.do();
manager2.do();
}
Everything in method doStuff() will all be committed together, unless there is an exception thrown, in which case it will all be rolled back.
I would recommend you to use Programmatic approach for this case.
Programmatic transaction management: This means that you have manage the transaction with the help of programming. That gives you extreme flexibility, but it is difficult to maintain.
Vs
Declarative transaction management: This means you separate transaction management from the business code. You only use annotations or XML based configuration to manage the transactions.
Perhaps, alternative way for your questions by Programmatic transaction management.
Eg.
/** DataSourceTransactionManager */
#Autowired
private PlatformTransactionManager txManager;
public void yourMethod() {
try {
// Start a manual transaction.
TransactionStatus status = getTransactionStatus();
your code...
.....
.....
//your condition
txManager.commit(status);
//your condition
txManager.rollback(status);
} catch (YourException e) {
//your condition
txManager.rollback(status);
}
}
/**
* getTransactionStatus
*
* #return TransactionStatus
*/
private TransactionStatus getTransactionStatus() {
DefaultTransactionDefinition dtd = new DefaultTransactionDefinition();
dtd.setPropagationBehavior(TransactionDefinition.PROPAGATION_REQUIRES_NEW);
dtd.setIsolationLevel(TransactionDefinition.ISOLATION_READ_COMMITTED);
dtd.setReadOnly(false);
return txManager.getTransaction(dtd);
}
Note: It does not mean you need to always use one approach like Programmatic transaction management. I prefer mixed approach. Please use easy way like Declarative transaction for simple database services, otherwise, just control with Programmatic transaction in your services will save your logic easily.
I want to create instances of a class which will have access to the underlying Embedded derby database and pass this class to each bundle binding to my database bundle using declarative services.
I have seen in the derby documentation that sharing one connection for multiple threads has many pitfalls. So I was thinking to create a connection for each instance of the class I am creating. Since I only want a very simple way to just create multiple connections and manage them, using "MiniConnectionPoolManager" here seems like a good option. The sample code for derby is shown below:
org.apache.derby.jdbc.EmbeddedConnectionPoolDataSource dataSource = new org.apache.derby.jdbc.EmbeddedConnectionPoolDataSource();
dataSource.setDatabaseName("c:/temp/testDB");
dataSource.setCreateDatabase("create");
MiniConnectionPoolManager poolMgr = new MiniConnectionPoolManager(dataSource, maxConnections);
...
Connection connection = poolMgr.getConnection();
...
connection.close();
But the documentation does not cover many things plus I am a beginner in using Database. My questions are:
When I am creating a new class that will need database connection to perform insert,update & other actions. Shall I pass the 'poolMgr' and call poolMgr.getConnection() from the newly created class?
When should I close this connection? I don't know for how long the bundle (user) will use the new class so shall I save the newly created connection in a private global variable and force the user to execute unregister class where I could then close the connection? Or shall I just close all connections when my database bundle is being deactivated.
Other suggestions are also appreciated to manage different classes accessing one database. Thank you in advance.
Edit:
The main class in my database bundle is always active as long as the application is running. It is the bundles requesting for an instance of a new class(performing database operation) that come and go. And also since it will be deployed in embedded system, I can only use small footprint applications.
You should get a connection from a connection pool when you need it and close the connection as soon as you can. It is the job of the connection pool to re-use connections, not yours.
In other words: Do not keep a connection alive until your consumer bundle is deactivated.
Connection pools normally implement DataSource interface, you should use the pools via it. In that case you can replace the pool implementation easily without changing your code. E.g:
#Component
public class MyComponent {
// Connection pool based DataSource
#Reference
DataSource dataSource;
public void myFunction() {
try (Connection c = dataSource.getConnection()) {
// Database operations
} catch (SQLException e) {
// TODO
}
}
}
When you find yourself repeating the same code many times (getting connection, catching SQLException), you can write a simple component that accepts functional interfaces. E.g.:
#Component
#Service
public class SQLHelper {
#Reference // This is a connection pool DataSource
private DataSource dataSource;
public <R> R execute(Callback<R> callback) {
try (Connection c = dataSource.getConnection()) {
return callback.call(c);
} catch (SQLException e) {
throw new UncheckedSQLException(e);
}
}
}
Your functional interface would look like this:
public interface Callback<R> {
R call(Connection connection);
}
And you would use it like this:
sqlHelper.execute((Connection c) -> {
// Do some stuff with the connection
});
Using transactions
If you want to use atomic transactions, I suggest that you should use org.apache.derby.jdbc.EmbeddedXADataSource together with org.apache.commons.dbcp.managed.BasicManagedDataSource from commons-dbcp. After that, you can handle transactions via JTA.
It is hard to use the JTA API directly. You should choose a library that helps you propagating transactions.
A small guide based on Declarative Services:
Install derby jar into your OSGi container
Install pax-derby bundle as well! By doing that, you will have a DataSourceFactory OSGi service
Install everit-dsf-bundle with its dependencies! You will see two new DS components. Create a configuration for the one called XADataSource via the webconsole! All configuration options have descriptions.
Install a JTA Transaction Manager into the OSGi container! You have several choices. I
Install everit-commons-dbcp-component with its dependencies! You will see two new DS components. Configure the Managed one in the webconsole and set the previously created XADataSource as the target! The transactional pool will take care of providing the same connection if you request-and-close connections whitin the scope of the same transaction.
normally use Aries Transaction Manager that embeds Geronimo TM.
Install everit-transaction-helper to your OSGi container! You will see a new OSGi service with the interface TransactionHelper (that is provided by a configurable DS component).
Now you have everything to write your code. Your component would similar to the following:
#Component
#Service
public class MyComponent {
#Reference
private DataSource dataSource;
#Reference
private TransactionHelper th;
public void myFunction() {
th.required(() -> {
try (Connection c = dataSource.getConnection()) {
// My SQL statements
} catch (SQLException e) {
// TODO
}
}
}
}
In case you do not need transaction handling, you can:
use the standard EmbeddedDataSource
use any non-transactional connection pool
skip the installation of the TransactionManager and TransactionHelper bundles
skip the usage of TransactionHelper from the code
A more complex guide (that also takes care of schema creation and uses OO based queries) is available at http://cookbook.everit.org/persistence/index.html.
Update
You do not have to get a connection for every SQL statement. You should get a connection, execute as many SQL statements that you can within a "moment" and than call close on the connection.
If you have to run three SQL statements right behind each other, you should request a connection, execute the three SQL statements and than call close on the connection
If you close the requested connection within the same function you requested it from the pool, you probably do things right. You might call other functions passing the connection as a parameter, but they should only use it to run SQL statements and than return.
You should not keep alive a connection and wait for another user action. That is the job of the connection pool. When you call close on a connection that is provided by a pool, the connection is not closed physically, but only retrieved to the pool.
You should keep the connection object in a local variable. If you use a member variable for your connection object, you should suspect that something is wrong with your code (the only exception is if you pass the Connection to an object that lives for a very short time and that object holds the connection in a member variable to have cleaner code).
Please note that if you use Java 6 or earlier, you should close the connection in a finally block to avoid unclosed connections.
MiniConnectionPoolManager might be a great solution for embedded devices as it is really "mini". The only issue is that it does not implement the DataSource interface so your business code shuold directly use the MiniCPM classes. By doing that, it will be much harder to switch to other Connection pool if you find a bug or you need a more complex pool later.
If you decide to use MiniCPM, I suggest that you should write a component that implements DataSource and delegates the getConnection() function to a MiniCPM instance. E.g.:
#Component
#Service
public class MiniCPMDataSourceComponent implements DataSource {
#Reference
protected ConnectionPoolDataSource cpDataSource;
private MiniConnectionPoolManager wrapped;
#Activate
public void activate() {
this.wrapped = new MiniConnectionPoolManager(cpDataSource);
}
#Override
public Connection getConnection() {
return wrapped.getConnection();
}
#Override
public Connection getConnection(String user, String password) {
throw new UnsupportedOperationException();
}
#Deactivate
public void deactivate() {
wrapped.dispose();
}
}
You can decorate this component with configuration possibilities like the max connection number and timeout (that is supported by MiniCPM). If you use the service that is provided by this component, you will be able to switch the connection pool without changing your business code. Also, your business bundle will not be wired directly to MiniCPM.
In my project I am using Spring and mybatis, with mybatis-spring gluing them together. I am using Spring's declarative #Transactional annotations around my service layers, which call to the mybatis mappers.
I would prefer that NONE of my mapper methods can be called without an active transaction. Instead, they quietly run in a transactionless context (or else they are starting a new transaction).
Is there a way to disable this behavior and make it act more like a MANDATORY propagation level?
The easiest way to do that is to create custom TransactionFactory which will check if there is active transaction and throw exception otherwise.
class MandatoryTransactionSpringManagedTransactionFactory extends SpringManagedTransactionFactory {
public Transaction newTransaction(DataSource dataSource, TransactionIsolationLevel level, boolean autoCommit) {
if (!TransactionSynchronizationManager.isActualTransactionActive()) {
throw new IllegalTransactionStateException(
"No existing transaction found during mapper invocation");
}
return super.newTransaction(dataSource);
}
}
It should be used to configure org.mybatis.spring.SqlSessionFactoryBean