An ISessionFactory is an expensive-to-create, threadsafe object intended to be shared by all application threads. An ISession is an inexpensive, non-threadsafe object that should be used once, for a single business process, and then discarded. For example, when using NHibernate in an ASP.NET application, pages could obtain an ISessionFactory using:

ISessionFactory sf = Global.SessionFactory;

Each call to a service method could create a new ISession, Flush() it, Commit() its transaction, Close() it and finally discard it. (The ISessionFactory may also be kept in a static Singleton helper variable.)

We use the NHibernate ITransaction API as discussed previously, a single Commit() of a NHibernate ITransaction flushes the state and commits any underlying database connection (with special handling of distributed transactions).

Ensure you understand the semantics of Flush(). Flushing synchronizes the persistent store with in-memory changes but not vice-versa. Note that for all NHibernate ADO.NET connections/transactions, the transaction isolation level for that connection applies to all operations executed by NHibernate!

The next few sections will discuss alternative approaches that utilize versioning to ensure transaction atomicity. These are considered "advanced" approaches to be used with care.

You should observe the following practices when creating NHibernate Sessions:

The application may concurrently access the same persistent state in two different units-of-work. However, an instance of a persistent class is never shared between two ISession instances. Hence there are two different notions of identity:

Then for objects attached to a particular Session, the two notions are equivalent. However, while the application might concurrently access the "same" (persistent identity) business object in two different sessions, the two instances will actually be "different" (CLR identity).

This approach leaves NHibernate and the database to worry about concurrency. The application never needs to synchronize on any business object, as long as it sticks to a single thread per ISession or object identity (within an ISession the application may safely use == to compare objects).

Many business processes require a whole series of interactions with the user interleaved with database accesses. In web and enterprise applications it is not acceptable for a database transaction to span a user interaction.

Maintaining isolation of business processes becomes the partial responsibility of the application tier, hence we call this process a long running application transaction. A single application transaction usually spans several database transactions. It will be atomic if only one of these database transactions (the last one) stores the updated data, all others simply read data.

The only approach that is consistent with high concurrency and high scalability is optimistic concurrency control with versioning. NHibernate provides for three possible approaches to writing application code that uses optimistic concurrency.

// foo is an instance loaded earlier by the Session
session.Reconnect();
transaction = session.BeginTransaction();
foo.Property = "bar";
session.Flush();
transaction.Commit();
session.Disconnect();

// foo is an instance loaded by a previous Session
foo.Property = "bar";
session = factory.OpenSession();
transaction = session.BeginTransaction();
session.SaveOrUpdate(foo);
session.Flush();
transaction.Commit();
session.Close();

// foo is an instance loaded by a previous Session
session = factory.OpenSession();
transaction = session.BeginTransaction();
int oldVersion = foo.Version;
session.Load( foo, foo.Key );
if ( oldVersion != foo.Version ) throw new StaleObjectStateException();
foo.Property = "bar";
session.Flush();
transaction.Commit();
session.close();

The first approach described above is to maintain a single ISession for a whole business process thats spans user think time. (For example, a servlet might keep an ISession in the user's HttpSession.) For performance reasons you should

before waiting for user activity. The method ISession.Disconnect() will disconnect the session from the ADO.NET connection and return the connection to the pool (unless you provided the connection).

ISession.Reconnect() obtains a new connection (or you may supply one) and restarts the session. After reconnection, to force a version check on data you aren't updating, you may call ISession.Lock() on any objects that might have been updated by another transaction. You don't need to lock any data that you are updating.

Heres an example:

ISessionFactory sessions;
IList<Foo> fooList;
Bar bar;
....
ISession s = sessions.OpenSession();
ITransaction tx = null;

try
{
    tx = s.BeginTransaction())

    fooList = s.CreateQuery(
    	"select foo from Eg.Foo foo where foo.Date = current date"
        // uses db2 date function
    ).List<Foo>();

    bar = new Bar();
    s.Save(bar);

    tx.Commit();
}
catch (Exception)
{
    if (tx != null) tx.Rollback();
    s.Close();
    throw;
}
s.Disconnect();

Later on:

s.Reconnect();

try
{
    tx = s.BeginTransaction();

    bar.FooTable = new HashMap();
    foreach (Foo foo in fooList)
    {
        s.Lock(foo, LockMode.Read);    //check that foo isn't stale
        bar.FooTable.Put( foo.Name, foo );
    }

    tx.Commit();
}
catch (Exception)
{
    if (tx != null) tx.Rollback();
    throw;
}
finally
{
    s.Close();
}

You can see from this how the relationship between ITransactions and ISessions is many-to-one, An ISession represents a conversation between the application and the database. The ITransaction breaks that conversation up into atomic units of work at the database level.

It is not intended that users spend much time worring about locking strategies. It's usually enough to specify an isolation level for the ADO.NET connections and then simply let the database do all the work. However, advanced users may sometimes wish to obtain exclusive pessimistic locks, or re-obtain locks at the start of a new transaction.

NHibernate will always use the locking mechanism of the database, never lock objects in memory!

The LockMode class defines the different lock levels that may be acquired by NHibernate. A lock is obtained by the following mechanisms:

The "explicit user request" is expressed in one of the following ways:

If ISession.Load() is called with Upgrade or UpgradeNoWait, and the requested object was not yet loaded by the session, the object is loaded using SELECT ... FOR UPDATE. If Load() is called for an object that is already loaded with a less restrictive lock than the one requested, NHibernate calls Lock() for that object.

ISession.Lock() performs a version number check if the specified lock mode is Read, Upgrade or UpgradeNoWait. (In the case of Upgrade or UpgradeNoWait, SELECT ... FOR UPDATE is used.)

If the database does not support the requested lock mode, NHibernate will use an appropriate alternate mode (instead of throwing an exception). This ensures that applications will be portable.

The legacy (1.0.x) behavior of NHibernate in regards to ADO.NET connection management was that a ISession would obtain a connection when it was first needed and then hold unto that connection until the session was closed. NHibernate introduced the notion of connection release modes to tell a session how to handle its ADO.NET connections. Note that the following discussion is pertinent only to connections provided through a configured IConnectionProvider; user-supplied connections are outside the breadth of this discussion. The different release modes are identified by the enumerated values of NHibernate.ConnectionReleaseMode:

The configuration parameter hibernate.connection.release_mode is used to specify which release mode to use. The possible values:

As of NHibernate, if your application manages transactions through .NET APIs such as System.Transactions library, ConnectionReleaseMode.AfterTransaction may cause NHibernate to open and close several connections during one transaction, leading to unnecessary overhead and transaction promotion from local to distributed. Specifying ConnectionReleaseMode.OnClose will revert to the legacy behavior and prevent this problem from occuring.