Potty Little Details

Just another WordPress.com weblog

Connection Pooling for the SQL Server .NET Data Provider

leave a comment »

Pooling connections can significantly enhance the performance and scalability of your application. The SQL Server .NET Data Provider provides connection pooling automatically for your ADO.NET client application. You can also supply several connection string modifiers to control connection pooling behavior (see the section “Controlling Connection Pooling with Connection String Keywords” later in this topic).

Pool Creation and Assignment

When a connection is opened, a connection pool is created based on an exact matching algorithm that associates the pool with the connection string in the connection. Each connection pool is associated with a distinct connection string. When a new connection is opened, if the connection string is not an exact match to an existing pool, a new pool is created.

In the following example, three new SqlConnection objects are created, but only two connection pools are required to manage them. Note that the first and second connection strings differ by the value assigned for Initial Catalog.

SqlConnection conn = new SqlConnection();

conn.ConnectionString = “Integrated Security=SSPI;Initial Catalog=northwind”;

conn.Open();

// Pool A is created.

SqlConnection conn = new SqlConnection();

conn.ConnectionString = “Integrated Security=SSPI;Initial Catalog=pubs”;

conn.Open();

// Pool B is created because the connection strings differ.

SqlConnection conn = new SqlConnection();

conn.ConnectionString = “Integrated Security=SSPI;Initial Catalog=northwind”;

conn.Open();

// The connection string matches pool A.

Once created, connection pools are not destroyed until the active process ends. Maintenance of inactive or empty pools involves minimal system overhead.

Connection Addition

A connection pool is created for each unique connection string. When a pool is created, multiple connection objects are created and added to the pool so that the minimum pool size requirement is satisfied. Connections are added to the pool as needed, up to the maximum pool size.

When a SqlConnection object is requested, it is obtained from the pool if a usable connection is available. To be usable, the connection must currently be unused, have a matching transaction context or not be associated with any transaction context, and have a valid link to the server.

If the maximum pool size has been reached and no usable connection is available, the request is queued. The object pooler satisfies these requests by reallocating connections as they are released back into the pool. If the time-out period (determined by the Connect Timeout connection string property) elapses before a connection object can be obtained, an error occurs.

CAUTION You must always close the Connection when you are finished using it. This can be done using either the Close or Dispose methods of the Connection object. Connections that are not explicitly closed are not added or returned to the pool.

Connection Removal

The object pooler will remove a connection from the pool if the connection lifetime has expired, or if the pooler detects that the connection with the server has been severed. Note that this can be detected only after attempting to communicate with the server. If a connection is found that is no longer connected to the server, it is marked as invalid. The object pooler periodically scans connection pools looking for objects that have been released to the pool and are marked as invalid. These connections are then permanently removed.

If a connection exists to a server that has disappeared, it is possible for this connection to be drawn from the pool even if the object pooler has not detected the severed connection and marked it as invalid. When this occurs, an exception is generated. However, you must still close the connection in order to release it back into the pool.

Transaction Support

Connections are drawn from the pool and assigned based on transaction context. The context of the requesting thread and the assigned connection must match. Therefore, each connection pool is actually subdivided into connections with no transaction context associated with them, and into N subdivisions that each contain connections with a particular transaction context.

When a connection is closed, it is released back into the pool and into the appropriate subdivision based on its transaction context. Therefore, you can close the connection without generating an error, even though a distributed transaction is still pending. This allows you to commit or abort the distributed transaction at a later time.

Controlling Connection Pooling with Connection String Keywords

The ConnectionString property of the SQLConnection object supports connection string key/value pairs that can be used to adjust the behavior of the connection pooling logic.

The following table describes the ConnectionString values you can use to adjust connection pooling behavior.

Name

Default

Description

Connection Lifetime

0

When a connection is returned to the pool, its creation time is compared with the current time, and the connection is destroyed if that time span (in seconds) exceeds the value specified by Connection Lifetime. This is useful in clustered configurations to force load balancing between a running server and a server just brought online.

A value of zero (0) will cause pooled connections to have the maximum time-out.

Connection Reset

‘true’

Determines whether the database connection is reset when being removed from the pool. For Microsoft SQL Server version 7.0, setting to false avoids making an additional server round trip when obtaining a connection, but you must be aware that the connection state, such as database context, is not being reset.

Enlist

‘true’

When true, the pooler automatically enlists the connection in the current transaction context of the creation thread if a transaction context exists.

Max Pool Size

100

The maximum number of connections allowed in the pool.

Min Pool Size

0

The minimum number of connections maintained in the pool.

Pooling

‘true’

When true, the connection is drawn from the appropriate pool, or if necessary, created and added to the appropriate pool.

Performance Counters for Connection Pooling

The SQL Server .NET Data Provider adds several performance counters that enable you to fine-tune connection pooling characteristics, detect intermittent problems related to failed connection attempts, and detect problems related to timed-out requests to your SQL Server.

The following table lists the connection pooling counters that can be accessed in Performance Monitor under the “.NET CLR Data” performance object.

Counter

Description

SqlClient: Current # of pooled and non pooled connections

Current number of connections, pooled or not.

SqlClient: Current # pooled connections

Current number of connections in all pools associated with the process.

SqlClient: Current # connection pools

Current number of pools associated with the process.

SqlClient: Peak # pooled connections

The highest number of connections in all pools since the process started. Note: this counter is only available when associated with a specific process instance. The _Global instance will always return 0.

SqlClient: Total # failed connects

The total number of connection open attempts that have failed for any reason.

Advertisements

Written by oneil

September 11, 2008 at 2:54 am

Posted in ADO dot net

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: