Skip Headers
Oracle® Call Interface Programmer's Guide,
11g Release 2 (11.2)

Part Number E10646-02
Go to Documentation Home
Home
Go to Book List
Book List
Go to Table of Contents
Contents
Go to Index
Index
Go to Master Index
Master Index
Go to Feedback page
Contact Us

Go to previous page
Previous
Go to next page
Next
View PDF

9 OCI Programming Advanced Topics

This chapter contains these topics:

Connection Pooling in OCI

Connection pooling is the use of a group (the pool) of reusable physical connections by several sessions, in order to balance loads. The management of the pool is done by OCI, not the application. Applications that can use connection pooling include middle-tier applications for Web application servers and e-mail servers.

A sample usage of this feature is in a Web application server connected to a back-end Oracle database. Suppose that a Web application server gets several concurrent requests for data from the database server. The application can create a pool (or a set of pools) in each environment during application initialization.

OCI Connection Pooling Concepts

Oracle has several transaction monitor capabilities such as the fine-grained management of database sessions and connections. This is done by separating the notion of database sessions (user handles) from connections (server handles). Using these OCI calls for session switching and session migration, it is possible for an application server or transaction monitor to multiplex several sessions over fewer physical connections, thus achieving a high degree of scalability by pooling of connections and back-end Oracle server processes.

The connection pool itself is normally configured with a shared pool of physical connections, translating to a back-end server pool containing an identical number of dedicated server processes.

The number of physical connections is less than the number of database sessions in use by the application. The number of physical connections and back-end server processes are also reduced by using connection pooling. Thus many more database sessions can be multiplexed.

Similarities and Differences from Shared Server

Connection pooling on the middle-tier is similar to what shared server offers on the back end. Connection pooling makes a dedicated server instance behave like a shared server instance by managing the session multiplexing logic on the middle tier.

The pooling of dedicated server processes including incoming connections into the dedicated server processes is controlled by the connection pool on the middle tier. The main difference between connection pooling and a shared server is that in the latter case, the connection from the client is normally to a dispatcher in the database instance. The dispatcher is responsible for directing the client request to an appropriate shared server. On the other hand, the physical connection from the connection pool is established directly from the middle-tier to the dedicated server process in the back-end server pool.

Connection pooling is beneficial only if the middle tier itself is multithreaded. Each thread can maintain a session to the database. The actual connections to the database are maintained by the connection pool and these connections (including the pool of dedicated database server processes) are shared among all the threads in the middle tier.

Stateless Sessions Versus Stateful Sessions

Stateless sessions are serially reusable across mid-tier threads. After a thread is done processing a database request on behalf of a three-tier user, the same database session can be reused for the purpose of a completely different request on behalf of a completely different three-tier user.

Stateful sessions to the database, on the other hand, are not serially reusable across mid-tier threads because they may have some particular state associated with a particular three-tier user. Examples of such state may be: open transactions, fetch state from a statement, PL/SQL package state, and so on. This makes the session not reusable for a different request for the duration that such state persists.

Note: Stateless sessions too may have open transactions, open statement fetch state, and so on. However, such a state persists for a relatively short duration (only during the processing of a particular three-tier request by a mid-tier thread) which allows the session to be serially reused for a different three-tier user (when such state is cleaned up).

Note: Stateless sessions are typically used in conjunction with Statement Caching.

What connection pooling offers is stateless connections and stateful sessions. Users who need to work with stateless sessions, see "Session Pooling in OCI".

Multiple Connection Pools

This advanced concept can be used for different database connections. Multiple connection pools can also be used when different priorities are assigned to users. Different service level guarantees can be implemented using connection pooling.

The following figure illustrates OCI connection pooling:

Figure 9-1 OCI Connection Pooling

Description of Figure 9-1 follows
Description of "Figure 9-1 OCI Connection Pooling"

Transparent Application Failover

Transaction Application Failover (TAF) is enabled for connection pooling. The concepts of TAF apply equally well with connections in the connection pool except that the BACKUP and PRECONNECT clauses should not be used in the connect string and do not work with connection pooling and TAF.

When a connection in the connection pool fails over, it uses the primary connect string itself to connect. Sessions failover when they use the pool for a database round trip after their instance failure. The listener would be configured to route it to a good instance if available, as is typical with service-based connect strings.

See Also:

Oracle Database Net Services Administrator's Guide, the chapter on Configuring Transparent Application Failover

OCI Calls for Connection Pooling

The steps in using connection pooling in your application are:

Allocate the Pool Handle

Connection pooling requires that the pool handle OCI_HTYPE_CPOOL be allocated by OCIHandleAlloc(). Multiple pools can be created for a given environment handle.

For a single connection pool, here is an allocation example:

OCICPool *poolhp;
OCIHandleAlloc((void *) envhp, (void **) &poolhp, OCI_HTYPE_CPOOL, 
                      (size_t) 0, (void **) 0));

Create the Connection Pool

The function OCIConnectionPoolCreate() initializes the connection pool handle. It has these IN parameters:

  • connMin, the minimum number of connections to be opened when the pool is created.

  • connIncr, the incremental number of connections to be opened when all the connections are busy and a call needs a connection. This increment is used only when the total number of open connections is less than the maximum number of connections that can be opened in that pool.

  • connMax, the maximum number of connections that can be opened in the pool. When the maximum number of connections are open in the pool, and all the connections are busy, if a call needs a connection, it will wait till it gets one. However, if the OCI_ATTR_CONN_NOWAIT attribute is set for the pool, an error is returned.

  • A poolUsername, and a poolPasswd, to allow user sessions to transparently migrate between connections in the pool.

  • In addition, an attribute OCI_ATTR_CONN_TIMEOUT, can be set to time out the connections in the pool. Connections idle for more than this time are terminated periodically, to maintain an optimum number of open connections. If this attribute is not set, then the connections are never timed out.

Note:

Shrinkage of the pool only occurs when there is a network round trip. If there are no operations, then the connections stay alive.

All the preceding attributes can be configured dynamically. So the application has the flexibility of reading the current load (number of open connections and number of busy connections) and tuning these attributes appropriately.

If the pool attributes (connMax, connMin, connIncr) are to be changed dynamically, OCIConnectionPoolCreate() must be called with mode set to OCI_CPOOL_REINITIALIZE.

The OUT parameters poolName and poolNameLen will contain values to be used in subsequent OCIServerAttach() and OCILogon2() calls in place of the database name and the database name length arguments.

There is no limit on the number of pools that can be created by an application. Middle tier applications can take advantage of this feature and create multiple pools to connect to the same server or to different servers, to balance the load based on the specific needs of the application.

Here is an example of this call:

OCIConnectionPoolCreate((OCIEnv *)envhp,
                   (OCIError *)errhp, (OCICPool *)poolhp,
                   &poolName, &poolNameLen,
                   (text *)database,strlen(database),
                   (ub4) conMin, (ub4) conMax, (ub4) conIncr,
                   (text *)pooluser,strlen(pooluser),
                   (text *)poolpasswd,strlen(poolpasswd),
                   OCI_DEFAULT));

Logon to the Database

The application will need to log on to the database for each thread, using one of the following interfaces.

  • OCILogon2()

    This is the simplest interface. Use this interface when you need a simple Connection Pool connection and do not need to alter any attributes of the session handle. This interface can also be used to make proxy connections to the database.

    Here is an example using OCILogon2():

    for (i = 0; i < MAXTHREADS; ++i) 
    { 
       OCILogon2(envhp, errhp, &svchp[i], "hr", 2, "hr", 2, poolName,
                 poolNameLen, OCI_LOGON2_CPOOL));
    
    }
    

    In order to use this interface to get a proxy connection, set the password parameter to NULL.

  • OCISessionGet()

    This is the recommended interface. It gives the user the additional option of using external authentication methods, such as certificates, distinguished name, and so on. OCISessionGet() is the recommended uniform function call to retrieve a session.

    Here is an example using OCISessionGet():

    for (i = 0; i < MAXTHREADS; ++i) 
    { 
            OCISessionGet(envhp, errhp, &svchp, authp,
                          (OraText *) poolName,
                          strlen(poolName), NULL, 0, NULL, NULL, NULL,
                          OCI_SESSGET_CPOOL)
     }
    
  • OCIServerAttach() and OCISessionBegin():

    Another interface can be used if the application must set any special attributes on the user session handle and server handle. For such a requirement, applications need to allocate all the handles (connection pool handle, server handles, session handles and service context handles).

    • Create the connection pool.

    • Call OCIServerAttach() with mode set to OCI_CPOOL.

    • Call OCISessionBegin() with mode set to OCI_DEFAULT.

The user should not set OCI_MIGRATE flag in the call to OCISessionBegin(), when the virtual server handle points to a connection pool (OCIServerAttach() called with mode set to OCI_CPOOL). Oracle supports passing this flag, OCI_MIGRATE, only for compatibility reasons. Do not use the OCI_MIGRATE flag, because the perception that the user gets when using a connection pool is of sessions having their own dedicated (virtual) connections which are transparently multiplexed onto real connections. Credentials can be set to OCI_CRED_RDBMS, OCI_CRED_EXT, or OCI_CRED_PROXY. If the credentials are set to OCI_CRED_EXT, no user name and no password need to be set on the session handle. If the credentials are set to OCI_CRED_PROXY, only user name must be set on the session handle. (no explicit primary session must be created and OCI_ATTR_MIGSESSION need not be set).

Connection pooling does the multiplexing of a virtual server handle over physical connections transparently, hence eliminating the need for users to do so. The user gets the feeling of a session having a dedicated (virtual) connection. Since the multiplexing is done transparently to the user, users must not attempt to multiplex sessions over the virtual server handles themselves. The concepts of session migration and session switching, which require explicit multiplexing at the user level, are defunct in the case of connection pooling and should not be used.

In an OCI program, the user should create (OCIServerAttach() with mode set to OCI_CPOOL) a unique virtual server handle for each session that is created using the connection pool. There should be a one-to-one mapping between virtual server handles and sessions.

Deal with SGA Limitations in Connection Pooling

With OCI_CPOOL mode (connection pooling), the session memory (UGA) in the back-end database will come out of the SGA. This may require some SGA tuning on the back-end database to have a larger SGA if your application consumes more session memory than the SGA can accommodate. The memory tuning requirements for the back-end database will be similar to configuring the LARGE POOL in case of a shared server back end except that the instance is still in dedicated mode.

See Also:

Oracle Database Performance Tuning Guide for more information, see the section on configuring Shared Server

If you are still running into the SGA limitation, you must consider:

  • Reducing the session memory consumption by having fewer open statements for each session

  • reducing the number of sessions in the back end by pooling sessions on the mid-tier or otherwise

  • turning off connection pooling

The application must avoid using dedicated database links on the back end with connection pooling.

If the back end is a dedicated server, effective connection pooling will not be possible because sessions using dedicated database links will be tied to a physical connection rendering that same connection unusable by other sessions. If your application uses dedicated database links and you do not see effective sharing of back-end processes among your sessions, you must consider using shared database links.

See Also:

For more information about distributed databases, see the section on shared database links in Oracle Database Administrator's Guide

Logoff from the Database

Corresponding to the logon calls, these are the interfaces to use to log off from the database in connection pooling mode.

  • OCILogoff():

    If OCILogon2() was used to make the connection, OCILogoff() must be used to log off.

  • OCISessionRelease()

    If OCISessionGet() was called to make the connection, then OCISessionRelease() must be called to log off.

  • OCISessionEnd() and OCIServerDetach()

    If OCIServerAttach() and OCISessionBegin() were called to make the connection and start up the session, then OCISessionEnd() must be called to end the session and OCIServerDetach() must be called to release the connection.

Destroy the Connection Pool

Use OCIConnectionPoolDestroy() to destroy the connection pool.

Free the Pool Handle

The pool handle is freed using OCIHandleFree().

These last three actions are illustrated in this code fragment:

for (i = 0; i < MAXTHREADS; ++i)
  {
    checkerr(errhp, OCILogoff((void *) svchp[i], errhp));
  }
  checkerr(errhp, OCIConnectionPoolDestroy(poolhp, errhp, OCI_DEFAULT));
  checkerr(errhp, OCIHandleFree((void *)poolhp, OCI_HTYPE_CPOOL));

Examples of OCI Connection Pooling

Examples of connection pooling in tested complete programs can be found in cdemocp.c and cdemocpproxy.c in directory demo.

Session Pooling in OCI

Session pooling means that the application will create and maintain a group of stateless sessions to the database. These sessions will be handed over to thin clients as requested. If no sessions are available, a new one may be created. When the client is done with the session, the client will release it to the pool. Thus, the number of sessions in the pool can increase dynamically.

Some of the sessions in the pool may be 'tagged' with certain properties. For instance, a user may request for a default session, set certain attributes on it, then label it or 'tag' it and return in to the pool. That user, or some other user, can require a session with the same attributes, and thus request for a session with the same tag. There may be several sessions in the pool with the same tag. The 'tag' on a session can be changed or reset.

Proxy sessions, too, can be created and maintained through this interface.

The behavior of the application when no free sessions are available and the pool has reached it's maximum size, will depend on certain attributes. A new session may be created or an error returned, or the thread may just block and wait for a session to become free.

The main benefit of this type of pooling will be performance. Making a connection to the database is a time-consuming activity, especially when the database is remote. Thus, instead of a client spending time connecting to the server, authenticating its credentials, and then receiving a valid session, it can just pick one from the pool.

Functionality of OCI Session Pooling

Session pooling has the following features:

  • Create, maintain and manage a pool of stateless sessions transparently.

  • Provide an interface for the application to create a pool and specify the minimum, increment and maximum number of sessions in the pool.

  • Provide an interface for the user to obtain and release a default or 'tagged' session to the pool. A 'tagged' session is one with certain client-defined properties.

  • Allow the application to dynamically change the number of minimum and maximum number of sessions.

  • Provide a mechanism to always maintain an optimum number of open sessions, by closing sessions that have been idle for very long, and creating sessions when required.

  • Allow for session pooling with authentication.

Homogeneous and Heterogeneous Session Pools

A session pool can be either homogeneous or heterogeneous. Homogeneous session pooling means that sessions in the pool are alike with respect to authentication (have the same user name and password and privileges). Heterogeneous session pooling means that you must provide authentication information because the sessions can have different security attributes and privileges.

Using Tags in Session Pools

The tags provide a way for users to customize sessions in the pool. A client may get a default or untagged session from a pool, set certain attributes on the session (such as NLS settings), and return the session to the pool, labeling it with an appropriate tag in the OCISessionRelease() call.

The user, or some other user, may request a session with the same tags in order to have a session with the same attributes, and can do so by providing the same tag in the OCISessionGet() call.

See Also:

"OCISessionGet()" for a further discussion of tagging sessions.

OCI Handles for Session Pooling

These handle types are for session pooling:

OCISPool

This is the session pool handle. It is allocated using OCIHandleAlloc(). It must be passed to OCISessionPoolCreate(), and OCISessionPoolDestroy(). It has the attribute type OCI_HTYPE_SPOOL.

An example of the OCIHandleAlloc() call follows:

OCISPool *spoolhp;
OCIHandleAlloc((void *) envhp, (void **) &spoolhp, OCI_HTYPE_SPOOL, 
                        (size_t) 0, (void **) 0));

For an environment handle, multiple session pools can be created.

OCIAuthInfo

This is the authentication information handle. It is allocated using OCIHandleAlloc(). It is passed to OCISessionGet(). It supports all the attributes that are supported for user session handle. Please refer to user session handle attributes for more information. The authentication information handle has the attribute type OCI_HTYPE_AUTHINFO.

An example of the OCIHandleAlloc() call follows:

OCIAuthInfo *authp;
OCIHandleAlloc((void *) envhp, (void **) &authp, OCI_HTYPE_AUTHINFO, 
                      (size_t) 0, (void **) 0));

See Also:

Using OCI Session Pooling

The steps in writing a simple session pooling application which uses a user name and password are:

  • Allocate the session pool handle using OCIHandleAlloc() for an OCISPool handle. Multiple session pools can be created for an environment handle.

  • Create the session pool using OCISessionPoolCreate() with mode set to OCI_DEFAULT (for a new session pool). Refer to the function for a discussion of the other modes.

  • Loop for each thread. Create the thread with a function that does the following:

  • Allocate an authentication information handle of type OCIAuthInfo using OCIHandleAlloc().

  • Set the user name and password in the authentication information handle using OCIAttrSet().

  • Get a pooled session using OCISessionGet() with mode set to OCI_SESSGET_SPOOL.

  • Perform the transaction.

  • Commit or rollback the transactions.

  • Release the session (logoff) with OCISessionRelease().

  • Free the authentication information handle with OCIHandleFree().

  • End of the loop for each thread.

  • Destroy the session pool using OCISessionPoolDestroy().

OCI Calls for Session Pooling

Here are the usages for OCI calls for session pooling.

Allocate the Pool Handle

Session pooling requires that the pool handle OCI_HTYPE_SPOOL be allocated by calling OCIHandleAlloc().

Multiple pools can be created for a given environment handle. For a single session pool, here is an allocation example:

OCISPool *poolhp; 
OCIHandleAlloc((void *) envhp, (void **) &poolhp, OCI_HTYPE_SPOOL, (size_t) 0,
               (void **) 0));

Create the Pool Session

The function OCISessionPoolCreate() can be used to create the session pool. Here is an example of how to use this call:

OCISessionPoolCreate(envhp, errhp, poolhp, (OraText **)&poolName, 
              (ub4 *)&poolNameLen, database, 
              (ub4)strlen((const signed char *)database),
              sessMin, sessMax, sessIncr,
              (OraText *)appusername,
              (ub4)strlen((const signed char *)appusername),
              (OraText *)apppassword,
              (ub4)strlen((const signed char *)apppassword),
              OCI_DEFAULT);

Logon to the Database

These are the interfaces that can be used to logon to the database in session pooling mode.

  • OCILogon2()

    This is the simplest interface. However, it does not give the user the option of using tagging. Here is an example of how OCILogon2() can be used to log on to the database in session pooling mode:

    for (i = 0; i < MAXTHREADS; ++i) 
    { 
      OCILogon2(envhp, errhp, &svchp[i], "hr", 2, "hr", 2, poolName,
                poolNameLen, OCI_LOGON2_SPOOL));
    }
    
  • OCISessionGet()

    This is the recommended interface. It gives the user the option of using tagging to label sessions in the pool, and thus make it easier to retrieve specific sessions. An example of using OCISessionGet() follows. It is taken from cdemosp.c in the demo directory.:

    OCISessionGet(envhp, errhp, &svchp, authInfop,
                 (OraText *)database,strlen(database), tag,
                 strlen(tag), &retTag, &retTagLen, &found, 
                 OCI_SESSGET_SPOOL);
    

Logoff from the Database

Corresponding to the preceding logon calls, these are the interfaces to use to log off from the database in session pooling mode.

  • OCILogoff()

    If OCILogon2() was used to make the connection, OCILogoff() must be called to log off.

  • OCISessionRelease()

    If OCISessionGet() was called to make the connection, then OCISessionRelease() must be called to log off. Pending transactions are automatically committed.

Destroy the Session Pool

OCISessionPoolDestroy() must be called to destroy the session pool. Here is an example of how this call can be made:

OCISessionPoolDestroy(poolhp, errhp, OCI_DEFAULT);

Free the Pool Handle

OCIHandleFree() must be called to free the session pool handle. Here is how this call can be made:

OCIHandleFree((void *)poolhp, OCI_HTYPE_SPOOL);

Note:

Developers: you are advised to commit or rollback any open transaction before releasing the connection back to the pool. If this is not done, Oracle Database automatically commits any open transaction when the connection is released.

If an instance failure is detected while the session pool is being used, OCI tries to clean up the sessions to that instance.

Marking Sessions Explicitly as Stateful or Stateless

An application typically requires a specific database session for the duration of a unit of work. The session is said to be STATEFUL for this duration. At the end of this unit of work, if the application does not depend on retaining the specific session for subsequent units of work, the session is said to be STATELESS at this point.

As the application is aware of when a session transitions from being STATEFUL to STATELESS and vice-versa, the application can explicitly inform OCI regarding these transitions by using OCI_ATTR_SESSION_STATE.

This indication by the application or caller can allow OCI and the Oracle Database to potentially leverage this information for transparently performing certain scalability optimizations. For example, the session could be given to someone else who needs it when the application is not working on it. By the same token, the session could potentially be replaced by a different session when the application needs it again.

Here is an example of marking sessions in a code snippet:

wait_for_transaction_request();
do {
 
ub1 state;
 
/* mark  database session as STATEFUL  */
state = OCI_SESSION_STATEFUL;
checkerr(errhp, OCIAttrSet(usrhp, OCI_HTYPE_SESSION,
        &state, 0, OCI_ATTR_SESSION_STATE, errhp));
/* do database work comprising of one or more related calls to the database */
 
...
 
/* done with database work, mark session as stateless */
state = OCI_SESSION_STATELESS;
checkerr(errhp, OCIAttrSet(usrhp, OCI_HTYPE_SESSION,
         &state, 0,OCI_ATTR_SESSION_STATE, errhp));
 
wait_for_transaction_request();
 
} while(not _done); 

If a session was obtained from outside an OCI session pool, the session starts as being OCI_SESSION_STATEFUL and remains OCI_SESSION_STATEFUL throughout the life of the session unless the application explicitly changes it to OCI_SESSION_STATELESS at some point.

If a session is obtained from an OCI session pool, the session is by default marked as OCI_SESSION_STATEFUL when the first call is initiated on that session after getting it from the pool. The session is also by default marked as being OCI_SESSION_STATELESS when it is released to the pool. Hence, there is no need to set this attribute explicitly with an OCI session pool. OCI session pool does this transparently. Use this attribute only if you are not using OCI session pooling.

Example of OCI Session Pooling

Here is an example of session pooling in a tested complete program:

See Also:

cdemosp.c in directory demo.

Runtime Connection Load Balancing

Oracle Real Application Clusters (RAC) is a database option in which a single database is hosted by multiple instances on multiple nodes. RAC's shared disk method of clustering databases increases scalability because nodes can easily be added or freed to meet current needs and improve availability because if one node fails, another can assume its workload. It adds high availability and failover capacity to the database, because all instances have access to the whole database.

Balancing of work requests occurs at two different times: at connect time and at runtime. These are referred to as connect time load balancing provided by Oracle Net Services and runtime connection load balancing. For RAC environments, session pools use service metrics received from the RAC load balancing advisory through Fast Application Notification (FAN) events to balance application session requests. The work requests coming into the session pool can be distributed across the instances of RAC offering a service, using the current service performance.

See Also:

Connect time load balancing occurs when a session is first created by the application. It is necessary that the sessions that are part of the pool be well distributed across RAC instances, at the time they are first created. This ensures that sessions on each of the instances get a chance to execute work.

Runtime connection load balancing is basically routing work requests to sessions in a session pool that best serve the work. It comes into effect when selecting a session from an existing session pool and thus is a very frequent activity. For session pools that support services at one instance only, the first available session in the pool is adequate. When the pool supports services that span multiple instances, there is a need to distribute the work requests across instances so that the instances that are providing better service or have greater capacity get more requests.

Runtime connection load balancing is enabled by default in a release 11.1 or higher client talking to a server of 10.2 or higher. Setting the mode parameter to OCI_SPC_NO_RLB when calling OCISessionPoolCreate() disables runtime connection load balancing.

Receiving Load Balancing Advisory FAN Events

To receive the service metrics based on the service time the following are required:

  • RAC environment with Oracle Clusterware set up and enabled.

  • The application must have been linked with the threads library.

  • The OCI environment must be created in OCI_EVENTS and OCI_THREADED mode.

  • The service should be modified to set up its goal and the connection load balancing goal:

    EXEC DBMS_SERVICE.MODIFY_SERVICE("myService",
         DBMS_SERVICE.GOAL_SERVICE_TIME,
         clb_goal => dbms_service.clb_goal_short);
    

Database Resident Connection Pooling

Database Resident Connection Pooling (DRCP) provides a connection pool in the database server for typical Web application usage scenarios where the application acquires a database connection, works on it for a relatively short duration, and then releases it. DRCP pools "dedicated" servers, which is the equivalent of a server foreground and a database session combined and henceforth are referred to as the "pooled" servers. DRCP complements middle-tier connection pools that share connections between threads in a middle-tier process. In addition, DRCP enables sharing of database connections across middle-tier processes on the same middle-tier host and even across middle-tier hosts. This results in significant reduction in key database resources needed to support a large number of client connections, thereby reducing the database tier memory footprint and boosting the scalability of both middle-tier and database tiers. Having a pool of readily available servers also has the additional benefit of reducing the cost of creating and tearing down client connections.

DRCP is especially relevant for architectures with multi-process single threaded application servers (such as PHP/Apache) that cannot do middle-tier connection pooling. The database can still scale to tens of thousands of simultaneous connections with DRCP.

The "pooled" server model closely follows the "dedicated" model that is used to connect to Oracle by default. It just removes the overhead of dedicating a server to every connection when the server is required only for executing short transactions. DRCP allows a connection to acquire and then voluntarily relinquish the "pooled" server, for use by other "similar" connections. On being acquired by a connection, a "pooled" server essentially transforms into a "dedicated" server for that connection, until it is released back into the pool. Clients getting connections out of the database resident connection pool connect to an Oracle background known as the connection broker. The connection broker implements the pool functionality and multiplexes pooled servers among inbound connections from the client processes.

See Also:

Session Purity and Connection Class

There are two new settings that can be specified when obtaining a session using OCISessionGet().

Session Purity

Session purity specifies whether the application wants a "brand new" session or whether the application logic is set up to reuse a "pooled" session. OCISessionGet() has been enhanced to take in a purity setting of OCI_SESSGET_PURITY_NEW or OCI_SESSGET_PURITY_SELF. Alternatively, you can set OCI_ATTR_PURITY_NEW or OCI_ATTR_PURITY_SELF on the OCIAuthInfo handle prior to calling OCISessionGet(). Both these methods are equivalent.

Note:

When reusing a session from the pool, the NLS attributes of the server take precedence over that of the client.

For example, if the client has NLS_LANG set to french_france.us7ascii and if it is assigned a German session from the pool, the client session would be German.

Connection classes can be used to restrict sharing and to avoid the above problem.

Setting Session Purity Example

A connection pooling application needs a NEW session.

/* OCIAttrSet method */

ub4 purity = OCI_ATTR_PURITY_NEW;
OCIAttrSet (authInfop, OCI_HTYPE_AUTHINFO,  &purity, sizeof (purity),
            OCI_ATTR_PURITY, errhp);
OCISessionGet (envhp, errhp, &svchp, authInfop, poolName, poolNameLen, NULL, 0,
               NULL, NULL, NULL, OCI_SESSGET_SPOOL);
/* poolName is the name returned by OCISessionPoolCreate() */

/*  OCISessionGet mode method */
OCISessionGet (envhp, errhp, &svchp, authInfop, poolName, poolNameLen, NULL, 0,
               NULL, NULL, NULL, OCI_SESSGET_SPOOL | OCI_SESSGET_PURITY_NEW);
/* poolName is the name returned by OCISessionPoolCreate() */

Connection Class

Connection class defines a logical name for the type of connection required by the application. Sessions from the OCISessionPool cannot be shared by different users - A session first created for user HR is only given out to subsequent requests by user HR. The connection class setting allows for further separation between the sessions of a given user. The connection class setting lets different applications (connecting as the same database user) identify their sessions using a logical name that corresponds to the application. OCI then ensures that such sessions belonging to a particular connection class are not shared outside of the connection class.

OCI provides a new attribute on the OCIAuthInfo handle, OCI_ATTR_CONNECTION_CLASS that can be used to set the connection class. The connection class is a string attribute. OCI supports a maximum connection class length of 1024 bytes. The character '*' is a special character and is not allowed in the connection class name.

Setting the Connection Class Examples

An HRMS application needs sessions identified with the connection class HRMS.

OCISessionPoolCreate (envhp, errhp, spoolhp, &poolName, &poolNameLen, "HRDB",
    strlen("HRDB"), 0, 10, 1, "HR", strlen("HR"), "HR", strlen("HR"),
    OCI_SPC_HOMOGENEOUS);
 
OCIAttrSet (authInfop, OCI_HTYPE_AUTHINFO, "HRMS", strlen ("HRMS"),
    OCI_ATTR_CONNECTION_CLASS, errhp);
OCISessionGet (envhp, errhp, &svchp, authInfop, poolName, poolNameLen, NULL, 0,
    NULL, NULL, NULL, OCI_SESSGET_SPOOL);

A recruitment application needs sessions identified with the connection class RECMS.

OCISessionPoolCreate (envhp, errhp, spoolhp, &poolName, &poolNameLen, "HRDB",
    strlen("HRDB"), 0, 10, 1, "HR", strlen("HR"), "HR", strlen("HR"),
    OCI_SPC_HOMOGENEOUS);
 
OCIAttrSet (authInfop, OCI_HTYPE_AUTHINFO,  "RECMS", strlen("RECMS"),
    OCI_ATTR_CONNECTION_CLASS, errhp);
OCISessionGet (envhp, errhp, &svchp, authInfop, poolName, poolNameLen, NULL, 0,
    NULL, NULL, NULL, OCI_SESSGET_SPOOL);

Defaults for Session Purity and Connection Class

Table 9-1, "Defaults Used in Various Client Scenarios" illustrates the defaults used in various client scenarios:

Table 9-1 Defaults Used in Various Client Scenarios

Attribute or Setting Application Uses OCISessionGet() from Session Pool Other Connections are not Obtained from OCISessionPool

OCI_ATTR_PURITY

OCI_ATTR_PURITY_SELF

OCI_ATTR_PURITY_NEW

OCI_ATTR_CONNECTION_CLASS

OCI-generated globally unique name for each client-side session pool that is used as the default connection class for all connections in the OCISessionPool.

SHARED

Sharing of sessions

Sharing of sessions between threads requesting sessions from the OCISessionPool.

Sharing among all connections of a particular database using the default SHARED connection class.


Starting the Database Resident Connection Pool

The database administrator (DBA) must logon as SYSDBA and start the default pool, SYS_DEFAULT_CONNECTION_POOL, using DBMS_CONNECTION_POOL.START_POOL with the default settings.

For detailed information about configuring the pool, see Oracle Database Administrator's Guide.

Enabling Database Resident Connection Pooling

Any application can leverage the database resident connection pool by specifying :POOLED in the EZ Connect string or (SERVER=POOLED) in the TNS connect string.

Example of TNS Connect String for Leveraging DRCP

BOOKSDB = (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=oraclehost.company.com)
     (PORT=1521))(CONNECT_DATA = (SERVICE_NAME=books.company.com)(SERVER=POOLED)))

Example of EZ Connect String for Leveraging DRCP

oraclehost.company.com:1521/books.company.com:POOLED

Leveraging the Scalability of DRCP in an OCI Application

We consider three types of application scenarios and explain how each benefits from DRCP:

  1. Applications that do not use OCI Session Pooling and also do not specify any connection class or purity setting or specify a purity setting of NEW will simply get a brand new session from the DRCP. By the same token, when the application releases a connection back to the pool, the session is not shared with other instances of the same application by default. SQL*Plus is an example of a client that does not use OCI Session Pooling. It holds on to connections even when the connection is idle. As result, the pool server remains assigned to the client as long as the client session exists or as long as the client session does not log off. The application, however, does get the benefit of reusing an existing pooled server process.

  2. Applications that use OCISessionGet() outside of an OCISessionPool, and specify the connection class and set purity=SELF will be able to reuse both DRCP pooled server processes and sessions. However, following an OCISessionRelease(), OCI will terminate the connection to the connection broker. On the next OCISessionGet() call, the application will reconnect to the broker. Once it reconnects, the DRCP will assign a pooled server (and session) belonging to the connection class specified. Reconnecting, however, incurs the cost of connection establishment and re-authentication. Such applications achieve better sharing of DRCP resources (processes and sessions) but do not get the benefit of caching connections to the connection broker.

  3. Applications that use OCISessionPool APIs and specify the connection class and set purity=SELF make full use of the Database Resident Connection Pool functionality through reuse of both the pooled server process and the associated session and also getting the benefit of cached connections to the connection broker. Cached connections do not incur the cost of re-authentication on OCISessionGet().

OCISessionPool APIs have been extended to interoperate with the database resident connection pool. Sessions obtained using OCISessionGet() from the OCISessionPool are obtained from the DRCP and sessions released to the OCISessionPool using OCISessionRelease() release the sessions to DRCP. The OCISessionPool also transparently keeps connections to the connection broker cached, for better performance.

Features offered by the traditional client-side OCISessionPool, such as tagging, statement caching and TAF (Transparent application failover) are supported with DRCP also.

For information on client-side session pooling APIs:

Best Practices for Using DRCP

The steps to design an application that is able to leverage the full power of DRCP are very similar to the steps required for an application that uses the OCISessionPool.

The only additional step is that for best performance, when deployed to run with DRCP, it is recommended that the application specify an explicit connection class setting.

Multiple instances of the same application should specify the same connection class setting for best performance and enhanced sharing of DRCP resources. Take care to ensure that the different instances of the application can share database sessions.

Example of a Database Resident Connection Pooling Application

/* Assume all necessary handles are allocated */
 
/*   This middle-tier uses a single database user. Create a homogeneous
     client-side SP */
OCISessionPoolCreate (envhp, errhp, spoolhp, &poolName, &poolNameLen, "BOOKSDB",
    strlen("BOOKSDB"), 0, 10, 1, "SCOTT", strlen("SCOTT"), "password",
    strlen("password"), OCI_SPC_HOMOGENEOUS);
 
while (1)
{
   /* Process a client request */
   WaitForClientRequest();
   /* Application function */
 
   /* Set the Connection Class on the OCIAuthInfo handle that is passed as
      argument to OCISessionGet*/
   
   OCIAttrSet (authInfop, OCI_HTYPE_AUTHINFO,  "BOOKSTORE", strlen("BOOKSTORE"),
               OCI_ATTR_CONNECTION_CLASS, errhp);
 
   /* Purity need not be set as default is OCI_ATTR_PURITY_SELF for OCISessionPool
       connections */
 
   /* We can get a SCOTT session released by Mid-tier 2 */
   OCISessionGet(envhp, errhp, &svchp, authInfop, poolName, poolNameLen, NULL, 0,
                 NULL, NULL, NULL, OCI_SESSGET_SPOOL); 
 
   /* Database calls using the svchp obtained above  */
   OCIStmtExecute(...)
 
   /* This releases the pooled server on the database for reuse */
   OCISessionRelease (svchp, errhp, NULL, 0, OCI_DEFAULT);
}
 
/* Mid-tier is done - exiting */
OCISessionPoolDestroy (spoolhp, errhp, OCI_DEFAULT);

Deployment Example 1

The code immediately preceding is deployed in 10 middle tier hosts servicing the BOOKSTORE application. The database used is 11g (or earlier) database in dedicated server mode but with DRCP not enabled. The client side has 11g libraries. The connect string used is:

BOOKSDB = (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=oraclehost.company.com)
   (PORT=1521))(CONNECT_DATA = (SERVICE_NAME=books.company.com)))

In this case, the application gets dedicated server connections from the database.

Deployment Example 2

The code above is deployed in 10 middle tier hosts servicing the BOOKSTORE application. DRCP is enabled on the 11g database. Now all the middle tier processes can leverage the pooling capabilities offered by DRCP. The database resource requirement with DRCP will be much less than what would be required with dedicated server mode. Simply change the connect string to:

BOOKSDB = (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=oraclehost.company.com)
  (PORT=1521))(CONNECT_DATA = (SERVICE_NAME=books.company.com)(SERVER=POOLED)))

Compatibility and Migration

An OCI application linked with 11g client libraries will work unaltered against a

  • 11g database with DRCP disabled

  • pre-11g database server

  • 11g database server with DRCP enabled, when deployed with the DRCP connect string

Suitable clients benefit from enhanced scalability offered by DRCP provided that they are appropriately modified to use the OCISessionPool APIs with the connection class and purity settings as previously described.

Restrictions on Using Database Resident Connection Pooling

The following cannot be performed or used with pooled servers:

  • Shut down the database.

  • Stop the database resident connection pool.

  • Change the password for the connected user.

  • Use shared database links to connect to a database resident connection pool that is on a different instance.

  • Use Advanced Security Options (ASO) such as encryption, certificates, and so on.

  • Use migratable sessions on the server side, directly by using the OCI_MIGRATE option or indirectly using OCIConnectionPool().

  • Use Initial Client Roles.

  • Use Application Context attributes - OCI_ATTR_APPCTX_NAME, OCI_ATTR_APPCTX_VALUE, and so on.

Using DRCP with Custom Pools

DRCP is well integrated with OCI session pooling as described in "Database Resident Connection Pooling". Oracle highly recommends using OCI session pool as it is pre-integrated with DRCP, FAN and RLB.

However, if an application is built using its own custom connection pool (or if the application does not use any pooling at all, but has periods of time when the session is not used and the application does not depend on getting back the specific session for correctness), it can still integrate with DRCP. You can do this by leveraging the OCI_ATTR_SESSION_STATE attribute as described in "Marking Sessions Explicitly as Stateful or Stateless".

When an application flags a session as being OCI_SESSION_STATELESS, OCI leverages this session annotation to return the session transparently to the DRCP pool (when DRCP is enabled). Similarly, when the application indicates the session as being OCI_SESSION_STATEFUL, OCI leverages this changed session state annotation to transparently check out an appropriate session from the DRCP pool.

It is recommended that applications mark session state as promptly as possible to enable efficient utilization of the underlying database resources.

Note: Other DRCP attributes such as connection class and purity still need to be specified as described in detail earlier.

When to Use Connection Pooling, Session Pooling, or Neither

If database sessions are not reusable by mid-tier threads (that is, they are stateful) and the number of back-end server processes may cause scaling problems on the database, use OCI connection pooling.

If database sessions are reusable by mid-tier threads (that is, they are stateless) and the number of back-end server processes may cause scaling problems on the database, use OCI session pooling.

If database sessions are not reusable by mid-tier threads (that is, they are stateful) and the number of back-end server processes will never be large enough to potentially cause any scaling issue on the database, there is no need to use any pooling mechanism.

Note:

Having non-pooled sessions/connections will result in tearing down and re-creation of the database session/connection for every mid-tier user request. This can cause severe scaling problems on the database side and excessive latency for the fulfillment of the request. Hence, it is strongly recommended that one of the pooling strategies be adopted for mid-tier applications based on whether the database session is stateful or stateless.

In connection pooling, the pool element is a connection and in session pooling, the pool element is a session.

As with any pool, the pooled resource is locked by the application thread for a certain duration until the thread has done its job on the database and the resource is released. The resource is unavailable to other threads during its period of use. Hence, application developers need to be aware that any kind of pooling works effectively with relatively short tasks. If the application is performing a long running transaction for example, it may deny the pooled resource to other sharers for long periods of time leading to starvation. Hence, pooling should be used in conjunction with short tasks and the size of the pool should be sufficiently large to maintain the desired concurrency of transactions.

Also, note that with:

  1. OCI Connection Pool

    1. Connections to the database are pooled. Sessions are created and destroyed by the user. Each call to the database will pick up an appropriate available connection from the pool.

    2. The application is multiplexing several sessions over fewer physical connections to the database. The users can tune the pool configuration to achieve required concurrency.

    3. The life-time of the application sessions is independent of the life-time of the cached pooled connections.

  2. OCI Session Pool

    Sessions and connections are pooled by OCI. The application gets sessions from the pool and releases sessions back to the pool.

Functions for Session Creation

The choices are:

  1. OCILogon()

    This is the simplest way to get an OCI Session. The advantage is ease of obtaining an OCI service context. The disadvantage is that you cannot perform any advance OCI operations like session migration, proxy authentication, using a connection pool, or a session pool.

  2. OCILogon2()

    This includes the functionality of OCILogon() to get a session. This session may be a new one with a new underlying connection, or one that is started over a virtual connection from an existing connection pool, or one from an existing session pool. The mode parameter value that the function is called with determines its behavior.

    The user cannot modify the attributes (except OCI_ATTR_STMTCACHESIZE) of the service context returned by OCI.

    See Also:

    "OCILogon2()"
  3. OCISessionBegin()

    This supports all the various options of an OCI session such as proxy authentication, getting a session from a connection pool or a session pool, external credentials, and migratable sessions. This is the lowest level call where all handles are needed to be explicitly allocated and all attributes set, and OCIServerAttach() is to be called prior to this call.

  4. OCISessionGet()

    This is now the recommended method to get a session. This session may be a new one with a new underlying connection, or one that is started over a virtual connection from an existing connection pool, or one from an existing session pool. The mode parameter value that the function is called with determines its behavior. This works like OCILogon2() but additionally enables you to specify tags for obtaining specific sessions from the pool.

Choosing Between Different Types of OCI Sessions

The choices are:

  • Basic OCI Sessions

    This works by using user name and password over a dedicated OCI server handle. This is the no-pool mechanism. See earlier notes of when to use it.

    If authentication is obtained through external credentials, then user name or password is not required.

  • Session Pool Sessions

    These sessions are from the session pool cache. Some sessions may be tagged. These are stateless sessions. Each OCISessionGet() and OCISessionRelease() call gets and releases a session from the session cache. This saves the server from creating and destroying sessions.

    See the earlier notes on connection pool sessions versus session pooling sessions versus no-pooling sessions.

  • Connection Pool Sessions

    These are sessions created using OCISessionGet() and OCISessionBegin() calls from an OCI Connection Pool. There is no session cache as these are stateful sessions. Each call creates a new session and the user is responsible for terminating these sessions.

    The sessions are automatically migratable between the server handles of the connection pool. Each session can have user name and password or be a proxy session. See the earlier notes on connection pool sessions versus session pooling sessions versus no-pooling sessions.

  • Sessions Sharing a Server Handle

    You can multiplex several OCI sessions over a few physical connections. The application does this manually by having the same server handle for these multiple sessions. It is preferred to have the session multiplexing details be left to OCI by using the OCI Connection Pool APIs.

  • Proxy Sessions

    This is useful if the password of the client must be protected from the middle-tier. Proxy sessions can also be part of OCI Connection Pool and OCI Session Pool.

  • Migratable Sessions

With transaction handles being migratable, there should be no need for applications to use this older feature, in light of OCI Connection Pooling.

Statement Caching in OCI

Statement caching refers to the feature that provides and manages a cache of statements for each session. In the server, it means that cursors are ready to be used without the need to parse the statement again. Statement caching can be used with connection pooling and with session pooling, and will improve performance and scalability. It can be used without session pooling as well. OCI calls that implement statement caching are:

Statement Caching without Session Pooling in OCI

Users perform the usual OCI steps to logon. The call to obtain a session will have a mode that specifies whether statement caching is enabled for the session. Initially the statement cache will be empty. Developers will try to find a statement in the cache using the statement text. If the statement exists the API will return a previously prepared statement handle, otherwise it will return a newly prepared statement handle.

The application developer can perform binds and defines and then simply execute and fetch the statement before returning the statement back to the cache. In the latter case, where the statement handle was not found, the developer will need to set different attributes on the handle in addition to the other steps.

OCIStmtPrepare2() will also take a mode which will determine if the developer wants a prepared statement handle or a null statement handle if the statement is not found in the cache.

The pseudo code will look like:

OCISessionBegin( userhp, ... OCI_STMT_CACHE)  ;
OCIAttrset(svchp, userhp, ...);  /* Set the user handle in the service context */
OCIStmtPrepare2(svchp, &stmthp, stmttext, key, ...);
OCIBindByPos(stmthp, ...);
OCIDefineByPos(stmthp, ...);
OCIStmtExecute(svchp, stmthp, ...);
OCIStmtFetch(svchp, ...);
OCIStmtRelease(stmthp, ...);
...

Statement Caching with Session Pooling in OCI

The concepts remain the same, except that the statement cache is enabled at the session pool layer rather than at the session layer.

The attribute OCI_ATTR_SPOOL_STMTCACHESIZE sets the default statement cache size for each of the sessions in the session pool to this value. It is set on the OCI_HTYPE_SPOOL handle. The statement cache size for a particular session in the pool can be overridden at any time by using OCI_ATTR_STMTCACHESIZE on that session. The value of OCI_ATTR_SPOOL_STMTCACHESIZE can be changed at any time. This attribute can be used to enable or disable statement caching at the pool level, after creation, just as attribute OCI_ATTR_STMTCACHESIZE (on the service context) is used to enable or disable statement caching at the session level. This change will be reflected on individual sessions in the pool, when they are handed to a user. Tagged sessions are an exception to this behavior. This is explained later.

Enabling or disabling of statement caching is allowed on individual pooled sessions similar to non-pooled sessions.

A user can enable statement caching on a session retrieved from a non-statement cached pool in an OCISessionGet() or OCILogon2() call by specifying OCI_SESSGET_STMTCACHE or OCI_LOGON2_STMTCACHE, respectively, in the mode argument.

When a user asks for a session from a session pool, the statement cache size for that session will default to that of the pool. This may also mean enabling or disabling statement caching in that session. For example, if a pooled session (session A) has statement caching enabled, and statement caching is turned off in the pool, and a user asks for a session, and session A is returned, then statement caching will be turned off in Session A. As another example, if Session A in a pool does not have statement caching enabled, and statement caching at the pool level is turned on, then before returning session A to a user, statement caching on Session A with size equal to that of the pool is turned on.

This will not hold true if a tagged session is asked for and retrieved. In this case, the size of the statement cache will not be changed. Consequently, it will not be turned on or off. Moreover, if the user specifies mode OCI_SESSGET_STMTCACHE in the OCISessionGet() call, this will be ignored if the session is tagged. In our earlier example, if Session A was tagged, then it is returned as is to the user.

Rules for Statement Caching in OCI

Here are some rules to follow:

  • Use the function OCIStmtPrepare2() instead of OCIStmtPrepare(). If you are using OCIStmtPrepare(), you are strongly urged not to use a statement handle across different service contexts. Doing so will raise an error if the statement has been obtained by OCIStmtPrepare2(). Migration of a statement handle to a new service context actually closes the cursor associated with the old session and therefore no sharing is achieved. Client-side sharing is also not obtained, because OCI will free all buffers associated with the old session when the statement handle is migrated.

  • You are urged to keep one service context for each session and use statement handles only for that service context. That will be the preferred and recommended model and usage.

  • A call to OCIStmtPrepare2(), even if the session does not have a statement cache, will also allocate the statement handle and therefore applications using only OCIStmtPrepare2() must not call OCIHandleAlloc() for the statement handle.

  • A call to the OCIStmtPrepare2() must be followed with OCIStmtRelease() after the user is done with the statement handle. If statement caching is used, this will release the statement to the cache. If statement caching is not used, the statement will be deallocated. Do not call OCIHandleFree() to free the memory.

  • If the call to OCIStmtPrepare2() is made with the OCI_PREP2_CACHE_SEARCHONLY mode and a NULL statement was returned (statement was not found), the subsequent call to OCIStmtRelease() is not required and must not be performed.

  • Do not call OCIStmtRelease() for a statement that was prepared using OCIStmtPrepare().

  • The statement cache has a maximum size (number of statements) which can be modified by an attribute on the service context, OCI_ATTR_STMTCACHESIZE. The default value is 20. This attribute can also be used to enable or disable statement caching for the session, pooled or non-pooled. If OCISessionBegin() is called without mode set as OCI_STMT_CACHE, then OCI_ATTR_STMTCACHESIZE can be set on the service context to a nonzero attribute to turn on statement caching. If statement caching is not turned on at the session pool level, OCISessionGet() will return a non-statement cache-enabled session. OCI_ATTR_STMTCACHESIZE can be used to turn the caching on. Similarly the same attribute can be used to turn off statement caching by setting the cache size to zero.

  • You can choose to tag a statement at the release time so that the next time you can request a statement of the same tag. The tag will be used to search the cache. An untagged statement (tag is NULL) is a special case of a tagged statement. Two statements are considered different if they only differ in their tags, or if one is untagged and the other is not.

    See Also:

    For information about the functions for statement caching:

Bind and Define Optimization in Statement Caching

In order to avoid repeated bind and define operations on statements in the cache by the application, the application can register an opaque context with a statement taken from the statement cache and register a callback function with the service context. The application data such as bind and define buffers can be enclosed in the opaque context. This context will be registered with the statement the first time it is taken from the cache. So when a statement is taken from the cache the second time and onwards, the application can reuse the bind and define buffers that it had registered with that statement. It is still the application's responsibility to manage the bind and defines. It can reuse both the bind and define data and the buffers or it can only change the data and reuse the buffers, or it can free and reallocate the buffers if the current size is not enough. In the last case, it needs to re-bind and re-define. In order to clean up the memory allocated by the application towards these bind and define buffers, the callback function is called during aging out of the statement or purging of the whole cache as part of session closure. The callback is called for every statement being purged. The application does the freeing of memory and any other cleanup required, inside the callback function. The pseudo-code is:

Get the statement using OCIStmtPrepare2(...)
 
Get the opaque context from the statement if exists
 
If opaque context doesn't exist
 
{
 
  Allocate fetch buffers,  do the OCIBindByPos, OCIDefineByPos etc.
 
  Enclose the buffer addresses inside a context and set the context and
  callback-function on the statement
 
}
Execute/Fetch using the statement and process the data in the fetch buffers. 
 
OCIStmtRelease() that statement
 
Next OCIStmtPrepare2()
 
OCIAttrGet() opaque application context from statement handle
 
Execute/Fetch using the statement and process the data in the fetch buffers.
 
OCIStmtRelease()
 
. . .
 
void callback_fn (context, statement, mode)
 
{
 
   /* mode= OCI_CBK_STMTCACHE_STMTPURGE means this got called when statement is
      aging out of the statement cache or if session is ended */
 
  <free the buffers in the context.>
 
}

OCI Statement Caching Code Example

Here is an example of statement caching:

See Also:

Please refer to cdemostc.c in directory demo for a working example of statement caching.

User-Defined Callback Functions in OCI

The Oracle Call Interface has the capability to execute user-specific code in addition to OCI calls. This functionality can be used for:

The OCI callback feature has been added by providing support for calling user code before or after executing the OCI calls. Functionality has also been provided to allow the user-defined code to be executed instead of executing the OCI code.

The user callback code can also be registered dynamically without modifying the source code of the application. The dynamic registration is implemented by loading up to five user-created dynamically linked libraries after the initialization of the environment handle during the OCIEnvCreate() call. These user-created libraries (such as Dynamic Link Libraries (DLLs) on Windows, or shared libraries on Solaris, register the user callbacks for the selected OCI calls transparently to the application.

Sample Application

For a listing of the complete demonstration programs that illustrate the OCI user callback feature, see Appendix B, "OCI Demonstration Programs".

Registering User Callbacks in OCI

An application can register user callback libraries with the OCIUserCallbackRegister() function. Callbacks are registered in the context of the environment handle. An application can retrieve information about callbacks registered with a handle with the OCIUserCallbackGet() function.

A user-defined callback is a subroutine that is registered against an OCI call and an environment handle. It can be specified to be either an entry callback, a replacement callback, or an exit callback.

  • If it is an entry callback, it is called when the program enters the OCI function.

  • Replacement callbacks are executed after entry callbacks. If the replacement callback returns a value of OCI_CONTINUE, then a subsequent replacement callback or the normal OCI-specific code is executed. If a replacement callback returns anything other than OCI_CONTINUE, subsequent replacement callbacks and the OCI code does not execute.

  • After a replacement callback returns something other than OCI_CONTINUE, or an OCI function successfully executes, program control transfers to the exit callback (if one is registered).

If a replacement or exit callback returns anything other than OCI_CONTINUE, then the return code from the callback is returned from the associated OCI call.

A user callback can return OCI_INVALID_HANDLE when either an invalid handle or an invalid context is passed to it.

Note:

If any callback returns anything other than OCI_CONTINUE, then that return code is passed to the subsequent callbacks. If a replacement or exit callback returns a return code other than OCI_CONTINUE, then the final (not OCI_CONTINUE) return code is returned from the OCI call.

OCIUserCallbackRegister

A user callback is registered using the OCIUserCallbackRegister() call.

Currently, OCIUserCallbackRegister() is only registered on the environment handle. The user's callback function of typedef OCIUserCallback is registered along with its context for the OCI call identified by the OCI function code, fcode. The type of the callback, whether entry, replacement, or exit, is specified by the when parameter.

For example, the stmtprep_entry_dyncbk_fn entry callback function and its context dynamic_context, are registered against the environment handle hndlp for the OCIStmtPrepare() call by calling the OCIUserCallbackRegister() function with the following parameters.

OCIUserCallbackRegister( hndlp, 
                         OCI_HTYPE_ENV, 
                         errh, 
                         stmtprep_entry_dyncbk_fn, 
                         dynamic_context, 
                         OCI_FNCODE_STMTPREPARE,
                         OCI_UCBTYPE_ENTRY
                         (OCIUcb*) NULL);

User Callback Function

The user callback function has to follow the following syntax:

typedef sword (*OCIUserCallback)
     (void *ctxp,      /* context for the user callback*/
      void *hndlp,     /* handle for the callback, env handle for now */
      ub4 type,         /* type of handlp, OCI_HTYPE_ENV for this release */
      ub4 fcode,        /* function code of the OCI call */
      ub1 when,         /* type of the callback, entry or exit */
      sword returnCode, /* OCI return code */
      ub4 *errnop,      /* Oracle error number */
      va_list arglist); /* parameters of the oci call */

In addition to the parameters described in the OCIUserCallbackRegister() call, the callback is called with the return code, errnop, and all the parameters of the original OCI as declared by the prototype definition.

The return code is always passed in as OCI_SUCCESS and *errnop is always passed in as 0 for the first entry callback. Note that *errnop refers to the content of errnop because errnop is an IN/OUT parameter.

If the callback does not want to change the OCI return code, then it must return OCI_CONTINUE, and the value returned in *errnop is ignored. If on the other hand, the callback returns any other return code than OCI_CONTINUE, the last returned return code becomes the return code for the call. At the this point, the value of *errnop returned is set in the error handle, or in the environment handle if the error information is returned in the environment handle because of the absence of the error handle for certain OCI calls such as OCIHandleAlloc().

For replacement callbacks, the returnCode is the non-OCI_CONTINUE return code from the previous callback or OCI call and *errnop is the value of the error number being returned in the error handle. This allows the subsequent callback to change the return code or error information if needed.

The processing of replacement callbacks is different in that if it returns anything other than OCI_CONTINUE, then subsequent replacement callbacks and OCI code is bypassed and processing jumps to the exit callbacks.

Note that if the replacement callbacks return OCI_CONTINUE to allow processing of OCI code, then the return code from entry callbacks is ignored.

All the original parameters of the OCI call are passed to the callback as variable parameters and the callback must retrieve them using the va_arg macros. The callback demonstration programs provide examples.

A null value can be registered to de-register a callback. That is, if the value of the callback (OCIUserCallback()) is NULL in the OCIUserCallbackRegister() call, then the user callback is de-registered.

When using the thread-safe mode, the OCI program acquires all mutexes before calling the user callbacks.

UserCallback Control Flow

This pseudocode describes the overall processing of a typical OCI call:

OCIXyzCall()
{
 Acquire mutexes on handles;
 retCode = OCI_SUCCESS;
 errno = 0;
 for all ENTRY callbacks do
  {
     
     EntryretCode = (*entryCallback)(..., retcode, &errno, ...);
     if (retCode != OCI_CONTINUE)
      {
         set errno in error handle or environment handle;
         retCode = EntryretCode;
       }
   }
  for all REPLACEMENT callbacks do
  {
   retCode = (*replacementCallback) (..., retcode, &errno, ...);
   if (retCode != OCI_CONTINUE)
      {
       set errno in error handle or environment handle
       goto executeEXITCallback;
       }
   }

   retCode = return code for XyzCall; /* normal processing of OCI call */

   errno = error number from error handle or env handle;

 executeExitCallback:
   for all EXIT callbacks do
   {
       exitRetCode = (*exitCallback)(..., retCode, &errno,...);
       if (exitRetCode != OCI_CONTINUE)
       {
           set errno in error handle or environment handle;
           retCode = exitRetCode;
       }
   }
    release mutexes;
    return retCode
}

UserCallback for OCIErrorGet()

If the callbacks are a total replacement of the OCI code, then they usually maintain their own error information in the call context and use that to return error information in bufp and errnop parameters of the replacement callback of the OCIErrorGet() call.

If on the other hand, the callbacks are either partially overriding OCI code, or just doing some other post processing, then they can use the exit callback to modify the error text and errnop parameters of the OCIErrorGet() by their own error message and error number. Note that the *errnop passed into the exit callback is the error number in the error or the environment handle.

Errors from Entry Callbacks

If an entry callback wants to return an error to the caller of the OCI call, then it must register a replacement or exit callback. This is because if the OCI code is executed, then the error code from the entry callback is ignored. Therefore the entry callback must pass the error to the replacement or exit callback through its own context.

Dynamic Callback Registrations

Because user callbacks are expected to be used for monitoring OCI behavior or to access other data sources, it is desirable that the registration of the callbacks be done transparently and non-intrusively. This is accomplished by loading user-created dynamically linked libraries at OCI initialization time. These dynamically linked libraries are called packages. The user-created packages register the user callbacks for the selected OCI calls. These callbacks can further register or de-register user callbacks as needed when receiving control at runtime.

A makefile (ociucb.mk on Solaris) is provided with the OCI demonstration programs to create the package. The exact naming and location of this package is operating system dependent. The source code for the package must provide code for special callbacks that are called at OCI initialization and environment creation times.

The loading of the package is controlled by setting an operating system environment variable, ORA_OCI_UCBPKG. This variable names the packages in a generic way. The packages must be located in the $ORACLE_HOME/lib directory.

Loading Multiple Packages

The ORA_OCI_UCBPKG variable can contain a semicolon separated list of package names. The packages are loaded in the order they are specified in the list.

For example, previously one specified the package as:

setenv ORA_OCI_UCBPKG mypkg

Now, you can still specify the package as earlier, but in addition multiple packages can be specified as:

setenv ORA_OCI_UCBPKG "mypkg;yourpkg;oraclepkg;sunpkg;msoftpkg"

All these packages are loaded in order. That is, mypkg is loaded first and msoftpkg is loaded last.

A maximum of five packages can be specified.

Note:

The sample makefile ociucb.mk creates ociucb.so.1.0 on a Solaris or ociucb.dll on a Windows system. To load the ociucb package, the environmental variable ORA_OCI_UCBPKG must be set to ociucb. On Solaris, if the package name ends with .so, OCIInitialize() fails. The package name must end with .so.1.0.

For further details about creating the dynamic link libraries, read the Makefiles provided in the demo directory for your operating system. For further information on user-defined callbacks, see your operating system-specific documentation on compiling and linking applications.

Package Format

Previously a package had to specify the source code for the OCIEnvCallback() function. Now the OCIEnvCallback() function is obsolete. Instead, the package source must provide two functions. The first function has to be named as packagename suffixed with the word Init. For example, if the package is named foo, then the source file (for example, but not necessarily, foo.c) must contain a fooInit() function with a call to OCISharedLibInit() function specified exactly as:

sword fooInit(metaCtx, libCtx, argfmt, argc, argv)
      void *     metaCtx;         /* The metacontext */
      void *     libCtx;          /* The context for this package. */
      ub4        argfmt;          /* package argument format */
      sword      argc;            /* package arg count*/
      void *     argv[];          /* package arguments */
{
  return  (OCISharedLibInit(metaCtx, libCtx, argfmt, argc, argv,
                            fooEnvCallback));
}

The last parameter of the OCISharedLibInit() function, fooEnvCallback(), in this case, is the name of the second function. It can be named anything, but by convention it can be named packagename suffixed with the word EnvCallback.

This function is a replacement for OCIEnvCallback(). Now all the dynamic user callbacks must be registered in this function. The function must be of type OCIEnvCallbackType, which is specified as:

typedef sword (*OCIEnvCallbackType)(OCIEnv *env, ub4 mode,
                                    size_t xtramem_sz, void *usrmemp,
                                    OCIUcb *ucbDesc);

When an environment handle is created, then this callback function is called at the very end. The env parameter is the newly created environment handle.

The mode, xtramem_sz, and usrmemp are the parameters passed to the OCIEnvCreate() call. The last parameter, ucbDesc, is a descriptor that is passed to the package. The package uses this descriptor to register the user callbacks as described later.

A sample ociucb.c file is provided in the demo directory. The makefile ociucb.mk is also provided (on Solaris) in the demo directory to create the package. Please note that this may be different on other operating systems. The demo directory also contains full user callback demo programs (cdemoucb.c, cdemoucbl.c,) illustrating this.

User Callback Chaining

User callbacks can both be registered statically in the application itself or dynamically at runtime in the DLLs. A mechanism is needed to allow the application to override a previously registered callback and then later invoke the overridden one in the newly registered callback to preserve the behavior intended by the dynamic registrations. This can result in chaining of user callbacks.

For this purpose, the OCIUserCallbackGet() function is provided to find out which function and context is registered for an OCI call.

Accessing Other Data Sources Through OCI

Because Oracle is the predominant database accessed, applications can take advantage of the OCI interface to access non-Oracle data by using the user callbacks to access them. This allows an application written in OCI to access Oracle data without any performance penalty. To access non-Oracle data sources, drivers can be written that access the non-Oracle data in user callbacks. Because OCI provides a very rich interface, there is usually a straightforward mapping of OCI calls to most data sources. This solution is better than writing applications for other middle layers such as ODBC that introduce performance penalties for all data sources. Using OCI does not incur any penalty for the common case of accessing Oracle data sources, and incurs the same penalty that ODBC does for non-Oracle data sources.

Restrictions on Callback Functions

There are certain restrictions on the usage of callback functions, including OCIEnvCallback():

  • A callback cannot call other OCI functions except OCIUserCallbackRegister(), OCIUserCallbackGet(), OCIHandleAlloc(), OCIHandleFree(). Even for these functions, if they are called in a user callback, then callbacks on them are not called to avoid recursion. For example, if OCIHandleFree() is called in the callback for OCILogoff(), then the callback for OCIHandleFree() is disabled during the execution of the callback for OCILogoff().

  • A callback cannot modify OCI data structures such as the environment or error handles.

  • A callback cannot be registered for OCIUserCallbackRegister() call itself, or for any of the following:

    • OCIUserCallbackGet()

    • OCIEnvCreate()

    • OCIInitialize()

    • OCIEnvInit()

Example of OCI Callbacks

For example, lets suppose that there are five packages each registering entry, replacement, and exit callbacks for the OCIStmtPrepare() call. That is, the ORA_OCI_UCBPKG variable is set as:

setenv ORA_OCI_UCBPKG "pkg1;pkg2;pkg3;pkg4;pkg5" 
 

In each package pkgN (where N can be 1 through 5), the pkgNInit() and PkgNEnvCallback() functions are specified as:

pkgNInit(void *metaCtx, void *libCtx, ub4 argfmt, sword argc, void **argv)
{
  return OCISharedLibInit(metaCtx, libCtx, argfmt, argc, argv, pkgNEnvCallback);
}

The pkgNEnvCallback() function registers the entry, replacement, and exit callbacks as:

pkgNEnvCallback(OCIEnv *env, ub4 mode, size_t xtramemsz,
                                void *usrmemp, OCIUcb *ucbDesc)
{
  OCIHandleAlloc((void *)env, (void **)&errh, OCI_HTYPE_ERROR, (size_t) 0,
        (void **)NULL);

  OCIUserCallbackRegister(env, OCI_HTYPE_ENV, errh, pkgN_entry_callback_fn,
        pkgNctx, OCI_FNCODE_STMTPREPARE, OCI_UCBTYPE_ENTRY, ucbDesc);

  OCIUserCallbackRegister(env, OCI_HTYPE_ENV, errh, pkgN_replace_callback_fn,
        pkgNctx, OCI_FNCODE_STMTPREPARE, OCI_UCBTYPE_REPLACE, ucbDesc);

  OCIUserCallbackRegister(env, OCI_HTYPE_ENV, errh, pkgN_exit_callback_fn,
        pkgNctx, OCI_FNCODE_STMTPREPARE, OCI_UCBTYPE_EXIT, ucbDesc);

  return OCI_CONTINUE;
}
 

Finally, in the source code for the application, user callbacks can be registered with the NULL ucbDesc as:

OCIUserCallbackRegister(env, OCI_HTYPE_ENV, errh, static_entry_callback_fn,
        pkgNctx, OCI_FNCODE_STMTPREPARE, OCI_UCBTYPE_ENTRY, (OCIUcb *)NULL);

  OCIUserCallbackRegister(env, OCI_HTYPE_ENV, errh, static_replace_callback_fn,
        pkgNctx, OCI_FNCODE_STMTPREPARE, OCI_UCBTYPE_REPLACE, (OCIUcb *)NULL);

  OCIUserCallbackRegister(env, OCI_HTYPE_ENV, errh, static_exit_callback_fn,
        pkgNctx, OCI_FNCODE_STMTPREPARE, OCI_UCBTYPE_EXIT, (OCIUcb *)NULL);
 

When the OCIStmtPrepare() call is executed, the callbacks are called in the following order:

static_entry_callback_fn() 
pkg1_entry_callback_fn() 
pkg2_entry_callback_fn() 
pkg3_entry_callback_fn() 
pkg4_entry_callback_fn() 
pkg5_entry_callback_fn() 
 
static_replace_callback_fn() 
 pkg1_replace_callback_fn() 
  pkg2_replace_callback_fn() 
   pkg3_replace_callback_fn() 
    pkg4_replace_callback_fn() 
     pkg5_replace_callback_fn() 
 
      OCI code for OCIStmtPrepare call 
 
pkg5_exit_callback_fn() 
pkg4_exit_callback_fn() 
pkg3_exit_callback_fn() 
pkg2_exit_callback_fn() 
pkg1_exit_callback_fn()

static_exit_callback_fn()

Note:

The exit callbacks are called in the reverse order of the entry and replacement callbacks

The entry and exit callbacks can return any return code and the processing continues to the next callback. However, if the replacement callback returns anything other than OCI_CONTINUE, then the next callback (or OCI code if it is the last replacement callback) in the chain is bypassed and processing jumps to the exit callback. For example, if pkg3_replace_callback_fn() returned OCI_SUCCESS, then pkg4_replace_callback_fn(), pkg5_replace_callback_fn(), and the OCI processing for the OCIStmtPrepare() call is bypassed. Instead, pkg5_exit_callback_fn() is executed next.

OCI Callbacks from External Procedures

There are several OCI functions that can be used as callbacks from external procedures.

See Also:

These functions are listed in Chapter 20, "OCI Cartridge Functions". For information about writing C subroutines that can be called from PL/SQL code, including a list of which OCI calls can be used, and some example code, refer to the Oracle Database Advanced Application Developer's Guide

Transparent Application Failover in OCI

Transparent Application Failover (TAF) is a client-side feature designed to minimize disruptions to end-user applications that occur when database connectivity fails because of instance or network failure. TAF can be implemented on a variety of system configurations including Real Application Clusters and Data Guard Physical Standby Databases. TAF can also be used after re-starting a single instance system, for example when repairs are made.

TAF can be configured to restore database sessions and, optionally, replay open queries. In previous versions, TAF with the SELECT failover option would be engaged only on the statement that was in use at the time of a failure. For example, if there were 10 statement handles in use by the application, and statement 7 was the failure-time statement (the statement in use when the failure happened), statements 1-6 and 8-10 would have to be re-executed after statement 7 was failed over using TAF.

Starting with 10g Release 2 (10.2), this has been improved. Now all statements that an application attempts to use after a failure will attempt failover. That is, an attempt to execute or fetch against other statements will engage TAF recovery just as for the failure-time statement. Notably, this means that subsequent statements may now succeed (whereas in the past they failed), or the application may receive errors corresponding to an attempted TAF recovery (like ORA-25401).

Note:

TAF is not supported for remote database links or for DML statements.

Configuring Transparent Application Failover

TAF can be configured on both the client-side and the server-side. If both are configured, server-side settings take precedence.

Configure TAF on the client side by including the FAILOVER_MODE parameter in the CONNECT_DATA portion of a connect descriptor.

See Also:

Oracle Database Net Services Reference for more information about client-side configuration of TAF (Connect Data Section).

Configure TAF on the server-side by modifying the target service with the DBMS_SERVICE.MODIFY_SERVICE packaged procedure.

An initial attempt at failover may not always be successful. The OCI provides a mechanism for retrying failover after an unsuccessful attempt.

See Also:

Oracle Database PL/SQL Packages and Types Reference for more information on the server-side configuration of TAF (DBMS_SERVICE).

Transparent Application Failover Callbacks in OCI

Because of the delay that can occur during failover, the application developer may want to inform the user that failover is in progress, and request that the user wait for notification that failover is complete. Additionally, the session on the initial instance may have received some ALTER SESSION commands. These ALTER SESSION commands will not be automatically replayed on the second instance. Consequently, the developer may wish to replay them on the second instance. OCIAttrSet() calls that affect the session must also be re-executed.

To accommodate these requirements, the application developer can register a failover callback function. In the event of failover, the callback function is invoked several times during the course of reestablishing the user's session.

The first call to the callback function occurs when the database first detects an instance connection loss. This callback is intended to allow the application to inform the user of an upcoming delay. If failover is successful, a second call to the callback function occurs when the connection is reestablished and usable.

Once the connection has been re-established, the client may wish to replay ALTER SESSION commands and inform the user that failover has happened. If failover is unsuccessful, then the callback is called to inform the application that failover will not take place. Additionally, the callback is called each time a user handle besides the primary handle is re-authenticated on the new connection. Since each user handle represents a server-side session, the client may wish to replay ALTER SESSION commands for that session.

See Also:

See "Handling OCI_FO_ERROR" for more information about this scenario

Failover Callback Structure and Parameters

The basic structure of a user-defined application failover callback function is as follows:

sb4 appfocallback_fn ( void       * svchp, 
                       void       * envhp, 
                       void       * fo_ctx, 
                       ub4        fo_type, 
                       ub4        fo_event );

An example is provided in the section "Failover Callback Example" for the following parameters:

svchp

The first parameter, svchp, is the service context handle. It is of type void *.

envhp

The second parameter, envhp, is the OCI environment handle. It is of type void *.

fo_ctx

The third parameter, fo_ctx, is a client context. It is a pointer to memory specified by the client. In this area the client can keep any necessary state or context. It is passed as a void *.

fo_type

The fourth parameter, fo_type, is the failover type. This lets the callback know what type of failover the client has requested. The usual values are:

  • OCI_FO_SESSION, which indicates that the user has requested only session failover.

  • OCI_FO_SELECT, which indicates that the user has requested select failover as well.

fo_event

The last parameter is the failover event. This indicates to the callback why it is being called. It has several possible values:

  • OCI_FO_BEGIN indicates that failover has detected a lost connection and failover is starting.

  • OCI_FO_END indicates successful completion of failover.

  • OCI_FO_ABORT indicates that failover was unsuccessful, and there is no option of retrying.

  • OCI_FO_ERROR also indicates that failover was unsuccessful, but it gives the application the opportunity to handle the error and retry failover.

  • OCI_FO_REAUTH indicates that you have multiple authentication handles and failover has occurred after the original authentication. It indicates that a user handle has been re-authenticated. To find out which, the application checks the OCI_ATTR_SESSION attribute of the service context handle (which is the first parameter).

Failover Callback Registration

For the failover callback to be used, it must be registered on the server context handle. This registration is done by creating a callback definition structure and setting the OCI_ATTR_FOCBK attribute of the server handle to this structure.

The callback definition structure must be of type OCIFocbkStruct. It has two fields: callback_function, which contains the address of the function to call, and fo_ctx which contains the address of the client context.

An example of callback registration is included as part of the example in the next section.

Failover Callback Example

The following code shows an example of a simple user-defined callback function definition, registration, and unregistration.

Part 1: Failover Callback Definition

sb4  callback_fn(svchp, envhp, fo_ctx, fo_type, fo_event)
void * svchp;
void * envhp;
void *fo_ctx;
ub4 fo_type;
ub4 fo_event;
{
switch (fo_event) 
   {
   case OCI_FO_BEGIN:
   {
     printf(" Failing Over ... Please stand by \n");
     printf(" Failover type was found to be %s \n",
                     ((fo_type==OCI_FO_SESSION) ? "SESSION" 
                     :(fo_type==OCI_FO_SELECT) ? "SELECT"
                     : "UNKNOWN!")); 
     printf(" Failover Context is :%s\n", 
                    (fo_ctx?(char *)fo_ctx:"NULL POINTER!"));
     break;
   }
   case OCI_FO_ABORT:
   {
     printf(" Failover stopped. Failover will not take place.\n");
     break;
   }
   case    OCI_FO_END:
   {
       printf(" Failover ended ...resuming services\n");
     break;
   }
   case OCI_FO_REAUTH:
   {
       printf(" Failed over user. Resuming services\n");
     break;
   }
   default:
   {
     printf("Bad Failover Event: %d.\n",  fo_event);
     break;
   }
   }
   return 0;
}

Part 2: Failover Callback Registration

int register_callback(srvh, errh)
void *srvh; /* the server handle */
OCIError *errh; /* the error handle */
{
  OCIFocbkStruct failover;                 /*  failover callback structure */
  /* allocate memory for context */
  if (!(failover.fo_ctx = (void *)malloc(strlen("my context.")+1)))
     return(1);
  /* initialize the context. */
  strcpy((char *)failover.fo_ctx, "my context.");
  failover.callback_function = &callback_fn;
  /* do the registration */
  if (OCIAttrSet(srvh, (ub4) OCI_HTYPE_SERVER,
                (void *) &failover, (ub4) 0,
                (ub4) OCI_ATTR_FOCBK, errh)  != OCI_SUCCESS)
     return(2);
  /* successful conclusion */
  return (0);
}

Part 3: Failover Callback Unregistration

OCIFocbkStruct failover;   /*  failover callback structure */
sword status;
 
  /* set the failover context to null */
  failover.fo_ctx = NULL; 
  /* set the failover callback to null */ 
  failover.callback_function = NULL; 
  /* un-register the callback */
  status = OCIAttrSet(srvhp, (ub4) OCI_HTYPE_SERVER,
                      (void *) &failover, (ub4) 0,
                      (ub4) OCI_ATTR_FOCBK, errhp);

Handling OCI_FO_ERROR

A failover attempt is not always successful. If the attempt fails, the callback function receives a value of OCI_FO_ABORT or OCI_FO_ERROR in the fo_event parameter. A value of OCI_FO_ABORT indicates that failover was unsuccessful, and no further failover attempts are possible. OCI_FO_ERROR, on the other hand, provides the callback function with the opportunity to handle the error in some way. For example, the callback may choose to wait a specified period of time and then indicate to the OCI library that it must reattempt failover.

Note:

This functionality is only available to applications linked with the 8.0.5 or later OCI libraries running against any Oracle server.

Failover does not work if a LOB column is part of the select list.

Consider the following time line of events:

Table 9-2 Time and Event

Time Event

T0

Database fails (failure lasts until T5).

T1

Failover triggered by user activity.

T2

User attempts to reconnect; attempt fails.

T3

Failover callback invoked with OCI_FO_ERROR.

T4

Failover callback enters predetermined sleep period.

T5

Database comes back up again.

T6

Failover callback triggers new failover attempt; it is successful.

T7

User successfully reconnects


The callback function triggers the new failover attempt by returning a value of OCI_FO_RETRY from the function.

The following example code shows a callback function which can be used to implement the failover strategy similar to the scenario described earlier. In this case the failover callback enters a loop in which it sleeps and then reattempts failover until it is successful:

/*--------------------------------------------------------------------*/
/* the user defined failover callback  */
/*--------------------------------------------------------------------*/
sb4  callback_fn(svchp, envhp, fo_ctx, fo_type, fo_event )
void * svchp;
void * envhp;
void *fo_ctx;
ub4 fo_type;
ub4 fo_event;
{
   OCIError *errhp;
   OCIHandleAlloc(envhp, (void **)&errhp, (ub4) OCI_HTYPE_ERROR,
              (size_t) 0, (void **) 0);
   switch (fo_event) 
   {
   case OCI_FO_BEGIN:
   {
     printf(" Failing Over ... Please stand by \n");
     printf(" Failover type was found to be %s \n",
            ((fo_type==OCI_FO_NONE) ? "NONE"
             :(fo_type==OCI_FO_SESSION) ? "SESSION" 
             :(fo_type==OCI_FO_SELECT) ? "SELECT"
             :(fo_type==OCI_FO_TXNAL) ? "TRANSACTION"
             : "UNKNOWN!")); 
     printf(" Failover Context is :%s\n", 
            (fo_ctx?(char *)fo_ctx:"NULL POINTER!"));
     break;
   }
   case OCI_FO_ABORT:
   {
     printf(" Failover aborted. Failover will not take place.\n");
     break;
   }
   case    OCI_FO_END:
   { 
       printf("\n Failover ended ...resuming services\n");
     break;
   }
   case OCI_FO_REAUTH:
   { 
       printf(" Failed over user. Resuming services\n");
     break;
   }
   case OCI_FO_ERROR:
   {
     /* all invocations of this can only generate one line. The newline
      * will be put at fo_end time.
      */
     printf(" Failover error gotten. Sleeping...");
     sleep(3);
     printf("Retrying. ");
     return (OCI_FO_RETRY);
     break;
   }
   default:
   {
     printf("Bad Failover Event: %d.\n",  fo_event);
     break;
   }
   }
   return 0;
}

HA Event Notification

Suppose an user employs a Web browser to log in to an application server that accesses a back-end database server. Failure of the database instance can result in a wait that can be up to in minutes in duration before the failure is known to the user. The ability to quickly detect failures of server instances, communicate this to the client, close connections, and clean up idle connections in connection pools is provided by HA event notification.

For High Availability clients connected to a Real Application Clusters (RAC) database, HA event notification can be used to provide a best-effort programmatic signal to the client in the event of a database failure. Client applications can register a callback on the environment handle to signal interest in this information. When a significant failure event occurs (which applies to a connection made by this client), the callback is invoked, with information concerning the event (the event payload), and a list of connections (server handles) that were disconnected as a result of the failure.

For example, consider a client application that has two connections to instance A and two connections to instance B of the same database. If instance A goes down, a notification of the event will be sent to the client, which will then disconnect the two connections to instance B, and invoke the registered callback. Note that if another instance, C, of the same database, goes down, the client will not be notified (since it will not affect any of the client's connections).

The HA event notification mechanism improves the response time of the application in the presence of failure. In the past, a failure would result in the connection being broken only after the TCP time out expired, which could take minutes. With HA event notification, standalone, connection pool, and session pool connections are automatically broken and cleaned up by OCI and the application callback is invoked within seconds of the failure event. If any of these server handles are TAF-enabled, failover will also automatically be engaged by OCI.

Applications must connect to a RAC instance to enable HA event notification. Furthermore, these applications must:

Then these applications can register a callback that is invoked whenever an HA event occurs.

OCIEvent Handle

The OCIEvent handle encapsulates the attributes from the event payload. OCI implicitly allocates this handle prior to calling the event callback, which can obtain the read-only attributes of the event by calling OCIAttrGet(). Memory associated with these attributes is only valid for the duration of the event callback.

OCI Failover for Connection and Session Pools

A connection pool in an instance of Real Application Clusters (RAC) consists of a pool of connections connected to different instances of the RAC. Upon receiving the node failure notification, all the connections connected to that particular instance should be cleaned up. For the connections that are in use, OCI has to close the connections: transparent application failover (TAF) occurs immediately and those connections will be reestablished. The connections that are idle and in the free list of the pool have to be purged, so that a bad connection is never returned back to the user from the pool.

To accommodate custom connection pools, OCI will provide a callback function that can be registered on the environment handle. If registered, this callback is invoked when an HA event occurs. Sessions pools are treated the same way as connection pools. Note that server handles from OCI connection pools or session pools will not be passed to the callback. Hence in some cases, the callback could be called with an empty list of connections.

OCI Failover for Independent Connections

No special handling is required for independent connections; all such connections that are connected to failed instances are immediately disconnected. For idle connections, TAF will be engaged to reestablish the connection when the connection is used on a subsequent OCI call. Connections that are in use at the time of the failure event will be broken out immediately, so that TAF can begin. Note that this applies for the "in-use" connections of connection and session pools also.

Event Callback

The event callback, of type OCIEventCallback, has the following signature:

void evtcallback_fn (void     *evtctx,
                     OCIEvent  *eventhp );

where evtctx is the client context and OCIEvent is an event handle which is opaque to the OCI library. The other input argument is eventhp, the event handle: that is, the attributes associated with an event.

If registered, this function will be called once for each event. In the case of RAC HA events, this callback will be invoked after the affected connections have been disconnected. The following environment handle attributes are used to register an event callback and context, respectively:

  • OCI_ATTR_EVTCBK is of datatype OCIEventCallback *. It is read-only.

  • OCI_ATTR_EVTCTX is of datatype void *. It is also read-only.

text *myctx = "dummy context"; /* dummy context passed to callback fn */
...
/* OCI_ATTR_EVTCBK and OCI_ATTR_EVTCTX are read-only. */
OCIAttrSet(envhp, (ub4) OCI_HTYPE_ENV, (void *) evtcallback_fn,
           (ub4) 0, (ub4) OCI_ATTR_EVTCBK, errhp);
OCIAttrSet(envhp, (ub4) OCI_HTYPE_ENV, (void *) myctx,
           (ub4) 0, (ub4) OCI_ATTR_EVTCTX, errhp);
...

Within the OCI Event callback, the list of affected server handles is encapsulated in the OCIEvent handle. For RAC HA DOWN events, client applications can iterate over a list of server handles that are affected by the event by using OCIAttrGet() with attribute types OCI_ATTR_HA_SVRFIRST and OCI_ATTR_HA_SVRNEXT:

OCIAttrGet(eventhp, OCI_HTYPE_EVENT, (void *)&srvhp, (ub4 *)0,
           OCI_ATTR_HA_SRVFIRST, errhp); 
/* or, */
OCIAttrGet(eventhp, OCI_HTYPE_EVENT, (void *)&srvhp, (ub4 *)0,
           OCI_ATTR_HA_SRVNEXT, errhp);

When called with attribute OCI_ATTR_HA_SRVFIRST, this function will retrieve the first server handle in the list of server handles affected. When called with attribute OCI_ATTR_HA_SRVNEXT, this function will retrieve the next server handle in the list. This function will return OCI_NO_DATA and srvhp will be a NULL pointer, when there are no more server handles to return.

srvhp is an output pointer to a server handle whose connection has been closed as a result of an HA event. errhp is an error handle to populate. The application returns an OCI_NO_DATA error when there are no more affected server handles to retrieve.

When retrieving the list of server handles that have been affected by an HA event, be aware that the connection has already been closed and many server handle attributes are no longer valid. Instead, use the user memory segment of the server handle to store any per-connection attributes required by the event notification callback. This memory remains valid until the server handle is freed.

Custom Pooling: Tagged Server Handles

The following features apply to custom pools:

  1. You can tag a server handle with its parent connection object if it is created on behalf of a custom pool. Use the "user memory" parameters of OCIHandleAlloc() to request that the server handle be allocated with a user memory segment. A pointer to the "user memory" segment is returned by OCIHandleAlloc().

  2. When an HA event occurs, and an affected server handle has been retrieved, there is a means to retrieve the server handle's tag information so appropriate cleanup can be performed. The attribute OCI_ATTR_USER_MEMORY is used to retrieve a pointer to a handle's user memory segment. OCI_ATTR_USER_MEMORY is valid for all user-allocated handles. If the handle was allocated with extra memory, this attribute will return a pointer to the user memory. A NULL pointer will be returned for those handles not allocated with extra memory. This attribute is read-only and is of datatype void*.

Note:

You are free to define the precise contents of the server handle's user memory segment to facilitate cleanup activities from within the HA event callback (or for other purposes if needed) because OCI does not write or read from this memory in any way. The user memory segment is freed along with the OCIHandleFree() call on the server handle.

Event Notification Example

sword retval;
OCIServer *srvhp;
struct myctx {
   void *parentConn_myctx;
   uword numval_myctx;
};
typedef struct myctx myctx; 
myctx  *myctxp;
/* Allocate a server handle with user memory - pre 10.2 functionality */
if (retval = OCIHandleAlloc(envhp, (void **)&srvhp, OCI_HTYPE_SERVER,
                            (size_t)sizeof(myctx), (void **)&myctxp)
/* handle error */
myctxp->parentConn_myctx = <parent connection reference>;
 
/* In an event callback function, retrieve the pointer to the user memory */
evtcallback_fn(void *evtctx, OCIEvent *eventhp)
{ 
  myctx *ctxp = (myctx *)evtctx;
  OCIServer *srvhp;
  OCIError *errhp;
  sb4       retcode;
  retcode = OCIAttrGet(eventhp, OCI_HTYPE_SERVER, &srvhp, (ub4 *)0,
                       OCI_ATTR_HA_SRVFIRST, errhp); 
  while (!retcode) /* OCIAttrGet will return OCI_NO_DATA if no more srvhp */ 
  {  
     OCIAttrGet((void *)srvhp, OCI_HTYPE_SERVER, (void *)&ctxp,
                (ub4)0, (ub4)OCI_ATTR_USER_MEMORY, errhp);
           /* Remove the server handle from the parent connection object */
     retcode = OCIAttrGet(eventhp, OCI_HTYPE_SERVER, &srvhp, (ub4 *)0,
                          OCI_ATTR_HA_SRVNEXT, errhp);
...
  }
...
}

Determining Transparent Application Failover (TAF) Capabilities

You can have the application adjust its behavior if a connection is or is not TAF-enabled. Use OCIAttrGet() as follows to find out if a server handle is TAF-enabled:

boolean taf_capable;
...
OCIAttrGet(srvhp, (ub4) OCI_HTYPE_SERVER, (void *) &taf_capable, 
           (ub4) sizeof(taf_capable), (ub4)OCI_ATTR_TAF_ENABLED, errhp);
...

where taf_capable is a boolean variable, which this call sets to TRUE if the server handle is TAF-enabled, and FALSE if not; srvhp is an input target server handle; OCI_ATTR_TAF_ENABLED is a an attribute which is a pointer to a boolean variable and is read-only; errhp is an input error handle.

OCI and Streams Advanced Queuing

The OCI provides an interface to Streams Advanced Queuing (Streams AQ) feature. Streams AQ provides message queuing as an integrated part of the Oracle server. Streams AQ provides this functionality by integrating the queuing system with the database, thereby creating a message-enabled database. By providing an integrated solution Streams AQ frees application developers to devote their efforts to their specific business logic rather than having to construct a messaging infrastructure.

Note:

In order to use Streams Advanced Queuing, you must be using the Enterprise Edition

See Also:

OCI Streams Advanced Queuing Functions

The OCI library includes several functions related to Streams Advanced Queuing:

  • OCIAQEnq()

  • OCIAQDeq()

  • OCIAQListen()

  • OCIAQListen2()

  • OCIAQEnqArray()

  • OCIAQDeqArray()

You can enqueue an array of messages to a single queue. The messages all share the same enqueue options, but each message in the array can have different message properties. You can also dequeue an array of messages from a single queue. For transaction group queues, you can dequeue all messages for a single transaction group using one call.

OCI Streams Advanced Queuing Descriptors

The following descriptors are used by OCI Streams AQ operations:

  • OCIAQEnqOptions

  • OCIAQDeqOptions

  • OCIAQMsgProperties

  • OCIAQAgent

You can allocate these descriptors with respect to the service handle using the standard OCIDescriptorAlloc() call. The following code shows examples of this:

OCIDescriptorAlloc(svch, &enqueue_options, OCI_DTYPE_AQENQ_OPTIONS, 0, 0 ); 
OCIDescriptorAlloc(svch, &dequeue_options, OCI_DTYPE_AQDEQ_OPTIONS, 0, 0 ); 
OCIDescriptorAlloc(svch, &message_properties, OCI_DTYPE_AQMSG_PROPERTIES, 0, 0);
OCIDescriptorAlloc(svch, &agent, OCI_DTYPE_AQAGENT, 0, 0 ); 

Each descriptor has a variety of attributes which can be set or read.

Streams Advanced Queuing in OCI Versus PL/SQL

The following tables compare functions, parameters, and options for OCI Streams AQ functions and descriptors, and PL/SQL AQ functions in the DBMS_AQ package.

Table 9-3 AQ Functions

PL/SQL Function OCI Function

DBMS_AQ.ENQUEUE

OCIAQEnq()

DBMS_AQ.DEQUEUE

OCIAQDeq()

DBMS_AQ.LISTEN

OCIAQListen(), OCIAQListen2()

DBMS_AQ.ENQUEUE_ARRAY

OCIAQEnqArray()

DBMS_AQ.DEQUEUE_ARRAY

OCIAQDeqArray()


The following table compares the parameters for the enqueue functions:

Table 9-4 Enqueue Parameters

DBMS_AQ.ENQUEUE Parameter OCIAQEnq() Parameter
queue_name
queue_name
enqueue_options
enqueue_options
message_properties
message_properties
payload
payload
msgid
msgid

-

Note: OCIAQEnq() also requires the following additional parameters: svch, errh, payload_tdo, payload_ind, and flags.


The following table compares the parameters for the dequeue functions:

Table 9-5 Dequeue Parameters

DBMS_AQ.DEQUEUE Parameter OCIAQDeq() Parameter
queue_name
queue_name
dequeue_options
dequeue_options
message_properties
message_properties
payload
payload
msgid
msgid

-

Note: OCIAQDeq() also requires the following additional parameters: svch, errh, dequeue_options, message_properties, payload_tdo, payload, payload_ind, and flags.


The following table compares parameters for the listen functions:

Table 9-6 Listen Parameters

DBMS_AQ.LISTEN Parameter OCIAQListen2() Parameter
agent_list
agent_list
wait
wait
agent
agent
listen_delivery_mode

lopts

-

Note: OCIAQListen2() also requires the following additional parameters: svchp, errhp, agent_list, num_agents, agent, lmops, and flags.


The following table compares parameters for the array enqueue functions:

Table 9-7 Array Enqueue Parameters

DBMS_AQ.ENQUEUE_ARRAY Parameter OCIAQEnqArray() Parameter
queue_name
queue_name
enqueue_options
enqopt
array_size
iters
message_properties_array
msgprop
payload_array
payload
msgid_array
msgid

-

Note: OCIAQEnqArray() also requires the following additional parameters: svch, errh, payload_tdo, payload_ind, ctxp, enqcbfp, and flags.


The following table compares parameters for the array dequeue functions:

Table 9-8 Array Dequeue Parameters

DBMS_AQ.DEQUEUE_ARRAY Parameter OCIAQDeqArray() Parameter
queue_name
queue_name
dequeue_options
deqopt
array_size
iters
message_properties_array
msgprop
payload_array
payload
msgid_array
msgid

-

Note: OCIAQDeqArray() also requires the following additional parameters: svch, errh, msgprop, payload_tdo, payload_ind, ctxp, deqcbfp, and flags.


The following table compares parameters for the agent attributes:

Table 9-9 Agent Parameters

PL/SQL Agent Parameter OCIAQAgent Attribute
name

OCI_ATTR_AGENT_NAME

address

OCI_ATTR_AGENT_ADDRESS

protocol

OCI_ATTR_AGENT_PROTOCOL


The following table compares parameters for the message properties:

Table 9-10 Message Properties

PL/SQL Message Property OCIAQMsgProperties Attribute
priority

OCI_ATTR_PRIORITY

delay

OCI_ATTR_DELAY

expiration

OCI_ATTR_EXPIRATION

correlation

OCI_ATTR_CORRELATION

attempts

OCI_ATTR_ATTEMPTS

recipient_list

OCI_ATTR_RECIPIENT_LIST

exception_queue

OCI_ATTR_EXCEPTION_QUEUE

enqueue_time

OCI_ATTR_ENQ_TIME

state

OCI_ATTR_MSG_STATE

sender_id

OCI_ATTR_SENDER_ID

transaction_group

OCI_ATTR_TRANSACTION_NO

original_msgid

OCI_ATTR_ORIGINAL_MSGID

delivery_mode

OCI_ATTR_MSG_DELIVERY_MODE


The following table compares enqueue option attributes:

Table 9-11 Enqueue Option Attributes

PL/SQL Enqueue Option OCIAQEnqOptions Attribute
visibility

OCI_ATTR_VISIBILITY

relative_msgid

OCI_ATTR_RELATIVE_MSGID

sequence_deviation

OCI_ATTR_SEQUENCE_DEVIATION

(deprecated)

transformation

OCI_ATTR_TRANSFORMATION

delivery_mode

OCI_ATTR_MSG_DELIVERY_MODE


The following table compares dequeue option attributes:

Table 9-12 Dequeue Option Attributes

PL/SQL Dequeue Option OCIAQDeqOptions Attribute
consumer_name

OCI_ATTR_CONSUMER_NAME

dequeue_mode

OCI_ATTR_DEQ_MODE

navigation

OCI_ATTR_NAVIGATION

visibility

OCI_ATTR_VISIBILITY

wait

OCI_ATTR_WAIT

msgid

OCI_ATTR_DEQ_MSGID

correlation

OCI_ATTR_CORRELATION

deq_condition

OCI_ATTR_DEQCOND

transformation

OCI_ATTR_TRANSFORMATION

delivery_mode

OCI_ATTR_MSG_DELIVERY_MODE


Note:

OCIAQEnq() returns the error ORA-25219 while specifying the enqueue option OCI_ATTR_SEQUENCE along with OCI_ATTR_RELATIVE_MSGID. This happens when enqueuing two messages. For the second message, enqueue options OCI_ATTR_SEQUENCE and OCI_ATTR_RELATIVE_MSGID is set to dequeue this message before the first one. It does not return an error if you do not specify the 'sequence' but, of course, the message is not dequeued before the relative message.

OCIAQEnq() does not return an error if OCI_ATTR_SEQUENCE attribute is not set, but the message is not dequeued before the message with relative message Id.

Buffered Messaging

Buffered messaging is a non-persistent messaging capability within Streams AQ, that was first available in Oracle Database 10g Release 2. Buffered messages reside in shared memory, and can be lost in the event of instance failure. Unlike persistent messages, no redo is written to disk. Buffered message enqueue and dequeue is much faster than persistent message operations. Since shared memory is limited, buffered messages may have to be spilled to disk. Flow control may be enabled to prevent applications from flooding the shared memory when the message consumers are slow or have stopped for some reason. The following functions are used for buffered messaging:

Enqueue Buffered Messaging Example

...
OCIAQMsgProperties  *msgprop;
OCIAQEnqueueOptions *enqopt;
message              msg;    /* message is an object type */
null_message         nmsg;   /* message indicator */
...
/* Allocate descriptors */
  OCIDescriptorAlloc(envhp, (void **)&enqopt, OCI_DTYPE_AQENQ_OPTIONS, 0,
                     (void **)0));
 
 OCIDescriptorAlloc(envhp, (void **)&msgprop,OCI_DTYPE_AQMSG_PROPERTIES, 0,
                    (void **)0));
 
/* Set delivery mode to buffered */
 dlvm = OCI_MSG_BUFFERED;
 OCIAttrSet(enqopt,  OCI_DTYPE_AQENQ_OPTIONS, (void *)&dlvm, sizeof(ub2),
            OCI_ATTR_MSG_DELIVERY_MODE, errhp);
/* Set visibility to Immediate (visibility must always be immediate for buffered
   messages) */
vis = OCI_ENQ_ON_COMMIT;
 
OCIAttrSet(enqopt, OCI_DTYPE_AQENQ_OPTIONS,(void *)&vis, sizeof(ub4),
           OCI_ATTR_VISIBILITY, errhp)
 
/* Message was an object type created earlier, msg_tdo is its Type
   Descriptor Object */
OCIAQEnq(svchp, errhp, "Test_Queue", enqopt, msgprop, msg_tdo, (void **)&mesg,
         (void **)&nmesg, (OCIRaw **)0, 0));
...

Dequeue Buffered Messaging Example

...
OCIAQMsgProperties  *msgprop;
OCIAQDequeueOptions *deqopt;
...
OCIDescriptorAlloc(envhp, (void **)&mprop, OCI_DTYPE_AQMSG_PROPERTIES, 0,
                   (void **)0));
OCIDescriptorAlloc(envhp, (void **)&deqopt, OCI_DTYPE_AQDEQ_OPTIONS, 0,
                   (void **)0);

/* Set visibility to Immediate (visibility must always be immediate for buffered
   message operations) */
vis = OCI_ENQ_ON_COMMIT;
OCIAttrSet(deqopt, OCI_DTYPE_AQDEQ_OPTIONS,(void *)&vis, sizeof(ub4),
           OCI_ATTR_VISIBILITY, errhp)
/* delivery mode is buffered */
dlvm  = OCI_MSG_BUFFERED;
OCIAttrSet(deqopt, OCI_DTYPE_AQDEQ_OPTIONS, (void *)&dlvm,  sizeof(ub2),
           OCI_ATTR_MSG_DELIVERY_MODE, errhp);
/* set the consumer for which to dequeue the message (this must be specified
   regardless of the type of message being dequeued.
*/
consumer = "FIRST_SUBSCRIBER";
OCIAttrSet(deqopt, OCI_DTYPE_AQDEQ_OPTIONS, (void *)consumer,
           (ub4)strlen((char*)consumer), OCI_ATTR_CONSUMER_NAME, errhp);
/* Dequeue the message but don't return the payload (to simplify the code
   snippet)
*/
OCIAQDeq(svchp, errhp,  "test_queue", deqopt, msgprop, msg_tdo, (void **)0,
         (void **)0, (OCIRaw**)0, 0);
...

Note:

Array operations are not supported for buffered messaging. Applications can use the OCIAQEnqArray() and OCIAQDeqArray() functions with array size set to 1.

Publish-Subscribe Notification in OCI

The publish-subscribe notification feature allows an OCI application to receive client notifications directly, register an e-mail address to which notifications can be sent, register a HTTP URL to which notifications can be posted, or register a PL/SQL procedure to be invoked on a notification. Figure 9-2, "Publish-Subscribe Model" illustrates the process.

Figure 9-2 Publish-Subscribe Model

Description of Figure 9-2 follows
Description of "Figure 9-2 Publish-Subscribe Model"

An OCI application can:

In all the preceding scenarios the notification can be received directly by the OCI application, or the notification can be sent to a pre-specified e-mail address, or it can be sent to a pre-defined HTTP URL, or a pre-specified database PL/SQL procedure can be invoked as a result of a notification.

Registered clients are notified asynchronously when events are triggered or on an explicit AQ enqueue. Clients do not need to be connected to a database.

See Also:

Publish-Subscribe Registration Functions in OCI

Registration can be done in two ways:

  • You register directly to the database. This way is simple and the registration will take effect immediately.

  • Open Registration. You register using LDAP, from which the database receives the registration request. This is useful when the client cannot have a database connection (the client wants to register for a database open event while the database is down), or if the client wants to register for the same event or events in multiple databases at one time.

Let us next consider these two alternative ways of registration.

Publish-Subscribe Register Directly to the Database

The following steps are required in an OCI application to register and receive notifications for events. It is assumed that the appropriate event trigger or AQ queue has been set up. The initialization parameter COMPATIBLE must be set to 8.1 or higher.

See Also:


Note:

The publish-subscribe feature is only available on multithreaded operating systems.
  1. Call OCIInitialize() with OCI_EVENTS mode to specify that the application is interested in registering for and receiving notifications. This starts a dedicated listening thread for notifications on the client.

  2. Call OCIHandleAlloc() with handle type OCI_HTYPE_SUBSCRIPTION to allocate a subscription handle.

  3. Call OCIAttrSet() to set the subscription handle attributes for:

    • OCI_ATTR_SUBSCR_NAME - subscription name

    • OCI_ATTR_SUBSCR_NAMESPACE - subscription namespace

    • OCI_ATTR_SUBSCR_HOSTADDR - environment handle attribute that sets the client IP (in either IPv4 or IPv6 format) to which notification is sent

      Oracle Database components and utilities support Internet Protocol version 6 (IPv6) addresses.

      See Also:

      "OCI_ATTR_SUBSCR_HOSTADDR", "OCI_ATTR_SUBSCR_IPADDR" and Oracle Database Net Services Administrator's Guide for more information about the IPv6 format for IP addresses
    • OCI_ATTR_SUBSCR_CALLBACK - notification callback

    • OCI_ATTR_SUBSCR_CTX - callback context

    • OCI_ATTR_SUBSCR_PAYLOAD - payload buffer for posting

    • OCI_ATTR_SUBSCR_RECPT - recipient name

    • OCI_ATTR_SUBSCR_RECPTPROTO - protocol to receive notification with

    • OCI_ATTR_SUBSCR_RECPTPRES - presentation to receive notification with

    • OCI_ATTR_SUBSCR_QOSFLAGS - QOS (quality of service) levels

    • OCI_ATTR_SUBSCR_TIMEOUT - Registration timeout interval in seconds. The default is 0 if a timeout is not set.

    • OCI_ATTR_SUBSCR_NTFN_GROUPING_CLASS - notification grouping class

      Notifications can be spaced out by using the grouping NTFN option with the following constants. A value supported for notification grouping class is:

      #define OCI_SUBSCR_NTFN_GROUPING_CLASS_TIME   1 /* time  */
      
    • OCI_ATTR_SUBSCR_NTFN_GROUPING_VALUE - notification grouping value in seconds

    • OCI_ATTR_SUBSCR_NTFN_GROUPING_TYPE - notification grouping type

      Supported values for notification grouping type:

      #define OCI_SUBSCR_NTFN_GROUPING_TYPE_SUMMARY 1  /* summary */
      #define OCI_SUBSCR_NTFN_GROUPING_TYPE_LAST    2  /* last */ 
      
    • OCI_ATTR_SUBSCR_NTFN_GROUPING_START_TIME - notification grouping start time

    • OCI_ATTR_SUBSCR_NTFN_GROUPING_REPEAT_COUNT - notification grouping repeat count

    • If OCI_SUBSCR_QOS_PURGE_ON_NTFN is set, the registration is purged on the first notification.

    • If OCI_SUBSCR_QOS_RELIABLE is set, notifications are persistent. Surviving instances of a RAC can be used to send and retrieve change notification messages, even after a node failure because invalidations associated with this registration are queued persistently into the database. If FALSE, then invalidations are enqueued into a fast in-memory queue. Note that this option describes the persistence of notifications and not the persistence of registrations. Registrations are automatically persistent by default.

    OCI_ATTR_SUBSCR_NAME, OCI_ATTR_SUBSCR_NAMESPACE and OCI_ATTR_SUBSCR_RECPTPROTO must be set before registering a subscription.

    If OCI_ATTR_SUBSCR_RECPTPROTO is set to OCI_SUBSCR_PROTO_OCI, then OCI_ATTR_SUBSCR_CALLBACK and OCI_ATTR_SUBSCR_CTX also need to be set.

    If OCI_ATTR_SUBSCR_RECPTPROTO is set to OCI_SUBSCR_PROTO_MAIL, OCI_SUBSCR_PROTO_SERVER, or OCI_SUBSCR_PROTO_HTTP, then OCI_ATTR_SUBSCR_RECPT also must be set.

    Setting OCI_ATTR_SUBSCR_CALLBACK and OCI_ATTR_SUBSCR_RECPT at the same time will cause an application error.

    OCI_ATTR_SUBSCR_PAYLOAD is required before posting to a subscription.

    See Also:

    "Subscription Handle Attributes" and "Creating the OCI Environment" for setting up the environment with mode = OCI_EVENTS | OCI_OBJECT. OCI_OBJECT is required for grouping notifications.
  4. The values of QOS, timeout interval, namespace, and port are set:

  5. If OCI_ATTR_SUBSCR_RECPTPROTO is set to OCI_SUBSCR_PROTO_OCI, then define the callback routine to be used with the subscription handle.

  6. If OCI_ATTR_SUBSCR_RECPTPROTO is set to OCI_SUBSCR_PROTO_SERVER, then define the PL/SQL procedure, to be invoked on notification, in the database.

  7. Call OCISubscriptionRegister() to register with the subscriptions. This call can register interest in several subscriptions at the same time.

Example of Setting QOS Levels

/* Set QOS levels */
ub4 qosflags = OCI_SUBSCR_QOS_PAYLOAD;
 
/* Set QOS flags in subscription handle */
(void) OCIAttrSet((dvoid *) subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
                 (dvoid *) &qosflags, (ub4) 0,
                 (ub4) OCI_ATTR_SUBSCR_QOSFLAGS, errhp);
 
/* Set notification grouping class */
ub4 ntfn_grouping_class = OCI_SUBSCR_NTFN_GROUPING_CLASS_TIME;
(void) OCIAttrSet((dvoid *) subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
                 (dvoid *) &ntfn_grouping_class, (ub4) 0,
                 (ub4) OCI_ATTR_SUBSCR_NTFN_GROUPING_CLASS, errhp);
 
/* Set notification grouping value of 10 minutes */
ub4 ntfn_grouping_value = 600;
(void) OCIAttrSet((dvoid *) subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
                 (dvoid *) &ntfn_grouping_value, (ub4) 0,
                 (ub4) OCI_ATTR_SUBSCR_NTFN_GROUPING_VALUE, errhp);
 
/* Set notification grouping type */
ub4 ntfn_grouping_type = OCI_SUBSCR_NTFN_GROUPING_TYPE_SUMMARY;
 
/* Set notification grouping type in subscription handle */
(void) OCIAttrSet((dvoid *) subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
                 (dvoid *) &ntfn_grouping_type, (ub4) 0,
                 (ub4) OCI_ATTR_SUBSCR_NTFN_GROUPING_TYPE, errhp);
 
/* Set namespace specific context */
(void) OCIAttrSet((dvoid *) subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
                 (dvoid *) NULL, (ub4) 0,
                 (ub4) OCI_ATTR_SUBSCR_NAMESPACE_CTX, errhp);

Open Registration for Publish-Subscribe

Prerequisites for this method are:

  • The compatibility of the database has to be 9.0 or higher.

  • LDAP_REGISTRATION_ENABLED must be set to TRUE. This can be done this way:

    ALTER SYSTEM SET LDAP_REGISTRATION_ENABLED=TRUE
    

    The default is FALSE.

  • LDAP_REG_SYNC_INTERVAL must be set to the time interval (in seconds) to refresh registrations from LDAP:

    ALTER SYSTEM SET LDAP_REG_SYNC_INTERVAL =  time_interval
    

    The default is 0, which means do not refresh.

  • To force a database refresh of LDAP registration information immediately:

    ALTER SYSTEM REFRESH LDAP_REGISTRATION
    

The steps in open registration using Oracle Enterprise Security Manager (OESM) are:

  1. In each enterprise domain, create enterprise role, ENTERPRISE_AQ_USER_ROLE.

  2. For each database in the enterprise domain, add global role GLOBAL_AQ_USER_ROLE to enterprise role ENTERPRISE_AQ_USER_ROLE.

  3. For each enterprise domain, add enterprise role ENTERPRISE_AQ_USER_ROLE to privilege group cn=OracleDBAQUsers, under cn=oraclecontext, under the administrative context.

  4. For each enterprise user that is authorized to register for events in the database, grant enterprise role ENTERPRISE_AQ_USER_ROLE.

Using OCI to Open Register with LDAP

  1. Call OCIInitialize() with mode set to OCI_EVENTS | OCI_USE_LDAP.

  2. Call OCIAttrSet() to set the following environment handle attributes for accessing LDAP:

    • OCI_ATTR_LDAP_HOST: the host name on which the LDAP server resides

    • OCI_ATTR_LDAP_PORT: the port on which the LDAP server is listening

    • OCI_ATTR_BIND_DN: the distinguished name to login to the LDAP server, usually the DN of the enterprise user

    • OCI_ATTR_LDAP_CRED: the credential used to authenticate the client, for example, the password for simple authentication (user name and password)

    • OCI_ATTR_WALL_LOC: for SSL authentication, the location of the client wallet

    • OCI_ATTR_LDAP_AUTH: the authentication method code

      See Also:

      "Environment Handle Attributes" for a complete list of authentication modes
    • OCI_ATTR_LDAP_CTX: the administrative context for Oracle in the LDAP server

  3. Call OCIHandleAlloc() with handle type OCI_HTYPE_SUBSCRIPTION, to allocate a subscription handle.

  4. Call OCIDescriptorAlloc() with descriptor type OCI_DTYPE_SRVDN, to allocate a server DN descriptor.

  5. Call OCIAttrSet() to set the server DN descriptor attributes for OCI_ATTR_SERVER_DN, the distinguished name of the database in which the client wants to receive notifications. OCIAttrSet() can be called multiple times for this attribute so that more than one database server is included in the registration

  6. Call OCIAttrSet() to set the subscription handle attributes for:

    • OCI_ATTR_SUBSCR_NAME - subscription name

    • OCI_ATTR_SUBSCR_NAMESPACE - subscription namespace

    • OCI_ATTR_SUBSCR_CALLBACK - notification callback

    • OCI_ATTR_SUBSCR_CTX - callback context

    • OCI_ATTR_SUBSCR_PAYLOAD - payload buffer for posting

    • OCI_ATTR_SUBSCR_RECPT - recipient name

    • OCI_ATTR_SUBSCR_RECPTPROTO - protocol to receive notification

    • OCI_ATTR_SUBSCR_RECPTRES - presentation to receive notification with

    • OCI_ATTR_SUBSCR_QOSFLAGS - QOS (quality of service) levels

    • OCI_ATTR_SUBSCR_TIMEOUT - Registration timeout interval in seconds. The default is 0 if a timeout is not set.

    • OCI_ATTR_SUBSCR_SERVER_DN - the descriptor handles you populated in step 5

  7. The values of QOS, timeout interval, namespace, and port are set:

  8. Call OCISubscriptionRegister() to register the subscriptions. The registration will take effect when the database accesses LDAP to pick up new registrations. The frequency of pick-ups is determined by the value of REG_SYNC_INTERVAL.

Setting QOS, Timeout Interval, Namespace, Client Address, and Port Number

You can set QOSFLAGS to the following QOS levels using OCIAttrSet():

  • OCI_SUBSCR_QOS_RELIABLE - Reliable notification persists across instance and database restarts. Reliability is of the server only and is only for persistent queues or buffered messages. This option describes the persistence of the notifications. Registrations are persistent by default.

  • OCI_SUBSCR_QOS_PURGE_ON_NTFN - Once received, purge registration on first notification. (Subscription is unregistered.)

/* Set QOS levels */
ub4 qosflags = OCI_SUBSCR_QOS_RELIABLE | OCI_SUBSCR_QOS_PURGE_ON_NTFN;

/* Set flags in subscription handle */
(void)OCIAttrSet((void *)subscrhp, (ub4)OCI_HTYPE_SUBSCRIPTION,
               (void *)&qosflags, (ub4)0, (ub4)OCI_ATTR_SUBSCR_QOSFLAGS, errhp);

/* Set auto-expiration after 30 seconds */
ub4 timeout = 30;
(void)OCIAttrSet((void *)subscrhp, (ub4)OCI_HTYPE_SUBSCRIPTION,
                 (void *)&timeout, (ub4)0, (ub4)OCI_ATTR_SUBSCR_TIMEOUT, errhp);

The registration is purged when the timeout is exceeded, and a notification is sent to the client, so that the client can invoke its callback and take any necessary action. In the case of client failure before the timeout, the registration is purged.

You can set the port number on the environment handle, which is important if the client is on a system behind a firewall that is able to receive notifications only on certain ports. Clients can specify the port for the listener thread before the first registration, using an attribute in the environment handle. The thread is started the first time OCISubscriptionRegister() is called. If available, this specified port number will be used. An error is returned if the client tries to start another thread on a different port using a different environment handle.

ub4 port = 1581;
(void)OCIAttrSet((void *)envhp, (ub4)OCI_HTYPE_ENV, (void *)&port, (ub4)0,
                 (ub4)OCI_ATTR_SUBSCR_PORTNO, errhp);

If instead, the port is determined automatically, you can get the port number at which the client thread is listening for notification by obtaining the attribute from the environment handle.

(void)OCIAttrGet((void *)subhp, (ub4)OCI_HTYPE_ENV, (void *)&port, (ub4)0, 
                 (ub4)OCI_ATTR_SUBSCR_PORTNO, errhp);

Example to set client address:

text ipaddr[16] = "10.177.246.40";
(void)(OCIAttrSet((dvoid *) envhp, (ub4) OCI_HTYPE_ENV,
       (dvoid *) ipaddr, (ub4) strlen((const char *)ipaddr),
       (ub4) OCI_ATTR_SUBSCR_IPADDR, errhp));

OCI Functions Used to Manage Publish-Subscribe Notification

The following functions are used to manage publish-subscribe notification.

Table 9-13 Publish-Subscribe Functions

Function Purpose

OCISubscriptionDisable()

Disables a subscription.

OCISubscriptionEnable()

Enables a subscription.

OCISubscriptionPost()

Posts a subscription.

OCISubscriptionRegister()

Registers a subscription.

OCISubscriptionUnRegister()

Unregisters a subscription.


Notification Callback in OCI

The client must register a notification callback that gets invoked when there is some activity on the subscription for which interest has been registered. In the AQ namespace, for instance, this occurs when a message of interest is enqueued.

This callback is typically set through the OCI_ATTR_SUBSCR_CALLBACK attribute of the subscription handle.

The callback must return a value of OCI_CONTINUE and adhere to the following specification:

typedef ub4 (*OCISubscriptionNotify) ( void            *pCtx,
                                       OCISubscription *pSubscrHp,
                                       void            *pPayload,
                                       ub4             iPayloadLen,
                                       void            *pDescriptor,
                                       ub4             iMode);

The parameters are described as follows:

pCtx (IN)

A user-defined context specified when the callback was registered.

pSubscrHp (IN)

The subscription handle specified when the callback was registered.

pPayload (IN)

The payload for this notification. For this release, only ub1 * (a sequence of bytes) for the payload is supported.

iPayloadLen (IN)

The length of the payload for this notification.

pDescriptor (IN)

The namespace-specific descriptor. Namespace-specific parameters can be extracted from this descriptor. The structure of this descriptor is opaque to the user and its type is dependent on the namespace.

The attributes of the descriptor are namespace-specific. For Advanced Queuing (AQ), the descriptor is OCI_DTYPE_AQNFY. For the AQ namespace, the count of notifications received in the group is provided in the notification descriptor. The attributes of pDescriptor are:

  • Notification Flag (regular = 0, timeout = 1, or grouping notification = 2) - OCI_ATTR_NFY_FLAGS

  • Queue Name - OCI_ATTR_QUEUE_NAME

  • Consumer Name - OCI_ATTR_CONSUMER_NAME

  • Message Id - OCI_ATTR_NFY_MSGID

  • Message Properties - OCI_ATTR_MSG_PROP

  • Count of Notifications Received In the Group - OCI_ATTR_AQ_NTFN_GROUPING_COUNT

  • The Group, an OCI Collection - OCI_ATTR_AQ_NTFN_GROUPING_MSGID_ARRAY

iMode (IN)

Call-specific mode. Valid value:

  • OCI_DEFAULT - executes the default call

Example of OCI Notification Callback for AQ Grouping

Here is an example of how to use the new AQ grouping notification attributes in a notification callback.

ub4 notifyCB1(void *ctx, OCISubscription *subscrhp, void *pay, ub4 payl,
              void *desc, ub4 mode)
{
 oratext            *subname;
 ub4                 size;
 OCIColl            *msgid_array = (OCIColl *)0;
 ub4                 msgid_cnt = 0;
 OCIRaw             *msgid;
 void              **msgid_ptr;
 sb4                 num_msgid = 0;
 void               *elemind = (void *)0;
 boolean             exist;
 ub2                 flags;
 oratext            *hexit = (oratext *)"0123456789ABCDEF";
 ub4                 i, j;
 
 /* get subscription name */
 OCIAttrGet(subscrhp, OCI_HTYPE_SUBSCRIPTION, (void *)&subname, &size,
            OCI_ATTR_SUBSCR_NAME,ctxptr->errhp);
 
 /* print subscripton name */
 printf("Got notification for %.*s\n", size, subname);
 fflush((FILE *)stdout);
 
 /* get the #ntfns received in this group */
 OCIAttrGet(desc, OCI_DTYPE_AQNFY, (void *)&msgid_cnt, &size,
            OCI_ATTR_AQ_NTFN_GROUPING_COUNT, ctxptr->errhp);
 
 /* get the group - collection of msgids */
 OCIAttrGet(desc, OCI_DTYPE_AQNFY, (void *)&msgid_array, &size,
            OCI_ATTR_AQ_NTFN_GROUPING_MSGID_ARRAY, ctxptr->errhp);
 
 /* get notification flag - regular, timeout, or grouping notification? */
 OCIAttrGet(desc, OCI_DTYPE_AQNFY, (void *)&flags, &size,
            OCI_ATTR_NFY_FLAGS, ctxptr->errhp);
 
 /* print notification flag */
 printf("Flag: %d\n", (int)flags);
 
 /* get group (collection) size */
 if (msgid_array)
   checkerr(ctxptr->errhp,
        OCICollSize(ctxptr->envhp, ctxptr->errhp,
        CONST OCIColl *) msgid_array, &num_msgid),
        "Inside notifyCB1-OCICollSize");
 else
   num_msgid =0;
 
 /* print group size */
 printf("Collection size: %d\n", num_msgid);
 
 /* print all msgids in the group */
 for(i = 0; i < num_msgid; i++)
 {
   ub4  rawSize;                                             /* raw size    */
   ub1 *rawPtr;                                              /* raw pointer */
     /* get msgid from group */
   checkerr(ctxptr->errhp,
        OCICollGetElem(ctxptr->envhp, ctxptr->errhp,
               (OCIColl *) msgid_array, i, &exist,
               (void **)(&msgid_ptr), &elemind),
        "Inside notifyCB1-OCICollGetElem");
   msgid = *msgid_ptr;
   rawSize = OCIRawSize(ctxptr->envhp, msgid);
   rawPtr = OCIRawPtr(ctxptr->envhp, msgid);
 
   /* print msgid size */
   printf("Msgid size: %d\n", rawSize);
 
   /* print msgid in hexadecimal format */
   for (j = 0; j < rawSize; j++)
   {                                           /* for each byte in the raw */
     printf("%c", hexit[(rawPtr[j] & 0xf0) >> 4]);
     printf("%c", hexit[(rawPtr[j] & 0x0f)]);
   }
   printf("\n");
 }
 
 /* print #ntfns received in group */
 printf("Notification Count: %d\n", msgid_cnt);
 printf("\n");
 printf("***********************************************************\n");
 fflush((FILE *)stdout);
 return 1;
}

Notification Procedure

The PL/SQL procedure that will be invoked when there is some activity on the subscription for which interest has been registered, has to be created in the database.

This procedure is typically set through the OCI_ATTR_SUBSCR_RECPT attribute of the subscription handle.

See Also:

Publish-Subscribe Direct Registration Example

This example shows how system events, client notification, and Advanced Queuing work together to implement publish subscription notification.

The following PL/SQL code creates all objects necessary to support a publish-subscribe mechanism under the user schema pubsub. In this code, the Agent snoop subscribes to messages that are published at logon events. Note that the user pubsub needs AQ_ADMINISTRATOR_ROLE and AQ_USER_ROLE privileges to use Advance Queuing functionality. The initialization parameter _SYSTEM_TRIG_ENABLED must be set to TRUE (the default) to enable triggers for system events. Connect as pubsub before running the following code:

----------------------------------------------------------
----create queue table for persistent multiple consumers
----------------------------------------------------------
---- Create or replace a queue table
begin
  DBMS_AQADM.CREATE_QUEUE_TABLE(
  QUEUE_TABLE=>'pubsub.raw_msg_table', 
  MULTIPLE_CONSUMERS => TRUE,
  QUEUE_PAYLOAD_TYPE =>'RAW',
  COMPATIBLE => '8.1.5');
end;
/
----------------------------------------------------------
---- Create a persistent queue for publishing messages
----------------------------------------------------------
---- Create a queue for logon events
begin
  DBMS_AQADM.CREATE_QUEUE(QUEUE_NAME=>'pubsub.logon',
  QUEUE_TABLE=>'pubsub.raw_msg_table',
  COMMENT=>'Q for error triggers');
end;
/
----------------------------------------------------------
---- Start the queue
----------------------------------------------------------
begin
  DBMS_AQADM.START_QUEUE('pubsub.logon');
end;
/
----------------------------------------------------------
---- define new_enqueue for convenience
----------------------------------------------------------
create or replace procedure new_enqueue(queue_name  in varchar2,
                                        payload  in raw ,
correlation in varchar2 := NULL,
exception_queue in varchar2 := NULL)
as
  enq_ct     dbms_aq.enqueue_options_t;
  msg_prop   dbms_aq.message_properties_t;
  enq_msgid  raw(16);
  userdata   raw(1000);
begin
  msg_prop.exception_queue := exception_queue;
  msg_prop.correlation := correlation;
  userdata := payload;
  DBMS_AQ.ENQUEUE(queue_name,enq_ct, msg_prop,userdata,enq_msgid);
end;
/
----------------------------------------------------------
---- add subscriber with rule based on current user name, 
---- using correlation_id
----------------------------------------------------------
declare
subscriber sys.aq$_agent;
begin
  subscriber := sys.aq$_agent('SNOOP', null, null);
  dbms_aqadm.add_subscriber(queue_name => 'pubsub.logon',
                            subscriber => subscriber,
                            rule => 'CORRID = ''ix'' ');
end;
/
----------------------------------------------------------
---- create a trigger on logon on database
----------------------------------------------------------
---- create trigger on after logon
create or replace trigger systrig2
   AFTER LOGON
   ON DATABASE
   begin
     new_enqueue('pubsub.logon', hextoraw('9999'), dbms_standard.login_user);
   end;
/

----------------------------------------------------------
---- create a PL/SQL callback for notification of logon 
---- of user 'ix' on database
----------------------------------------------------------
---- 
create or replace procedure plsqlnotifySnoop(
  context raw, reginfo sys.aq$_reg_info, descr sys.aq$_descriptor,
  payload raw, payloadl number)
as
begin
 dbms_output.put_line('Notification : User ix Logged on\n');
end;
/

After the subscriptions are created, the client must register for notification using callback functions. The following sample code performs the necessary steps for registration. The initial steps of allocating and initializing session handles are omitted here for sake of clarity.

...
static ub4 namespace = OCI_SUBSCR_NAMESPACE_AQ;

static OCISubscription *subscrhpSnoop = (OCISubscription *)0;
static OCISubscription *subscrhpSnoopMail = (OCISubscription *)0;
static OCISubscription *subscrhpSnoopServer = (OCISubscription *)0;

/* callback function for notification of logon of user 'ix' on database */

static ub4 notifySnoop(ctx, subscrhp, pay, payl, desc, mode)
    void *ctx;
    OCISubscription *subscrhp;
    void *pay;
    ub4 payl;
    void *desc;
    ub4 mode;
{
    printf("Notification : User ix Logged on\n");
  (void)OCIHandleFree((void *)subscrhpSnoop,
            (ub4) OCI_HTYPE_SUBSCRIPTION);
    return 1;
}

static void checkerr(errhp, status)
OCIError *errhp;
sword status;
{
  text errbuf[512];
  ub4 buflen;
  sb4 errcode;

  if (status == OCI_SUCCESS) return;

  switch (status)
  {
  case OCI_SUCCESS_WITH_INFO:
    printf("Error - OCI_SUCCESS_WITH_INFO\n");
    break;
  case OCI_NEED_DATA:
    printf("Error - OCI_NEED_DATA\n");
    break;
  case OCI_NO_DATA:
    printf("Error - OCI_NO_DATA\n");
    break;
  case OCI_ERROR:
    OCIErrorGet ((void *) errhp, (ub4) 1, (text *) NULL, &errcode,
            errbuf, (ub4) sizeof(errbuf), (ub4) OCI_HTYPE_ERROR);
    printf("Error - %s\n", errbuf);
    break;
  case OCI_INVALID_HANDLE:
    printf("Error - OCI_INVALID_HANDLE\n");
    break;
  case OCI_STILL_EXECUTING:
    printf("Error - OCI_STILL_EXECUTING\n");
    break;
  case OCI_CONTINUE:
    printf("Error - OCI_CONTINUE\n");
    break;
  default:
    printf("Error - %d\n", status);
    break;
  }
}

static void initSubscriptionHn (subscrhp,
                         subscriptionName,
                         func,
                         recpproto,
                         recpaddr,
                         recppres)
OCISubscription **subscrhp;
  char * subscriptionName;
  void * func;
  ub4 recpproto;
  char * recpaddr;
  ub4 recppres;
{
    /* allocate subscription handle */
    (void) OCIHandleAlloc((void *) envhp, (void **)subscrhp,
        (ub4) OCI_HTYPE_SUBSCRIPTION,
        (size_t) 0, (void **) 0);

    /* set subscription name in handle */
    (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
        (void *) subscriptionName,
        (ub4) strlen((char *)subscriptionName),
        (ub4) OCI_ATTR_SUBSCR_NAME, errhp);

    /* set callback function in handle */
    if (func)
      (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
          (void *) func, (ub4) 0,
          (ub4) OCI_ATTR_SUBSCR_CALLBACK, errhp);

    /* set context in handle */
    (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
        (void *) 0, (ub4) 0,
       (ub4) OCI_ATTR_SUBSCR_CTX, errhp);

    /* set namespace in handle */
    (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
        (void *) &namespace, (ub4) 0,
        (ub4) OCI_ATTR_SUBSCR_NAMESPACE, errhp);

    /* set receive with protocol in handle */
    (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
        (void *) &recpproto, (ub4) 0,
        (ub4) OCI_ATTR_SUBSCR_RECPTPROTO, errhp);

    /* set recipient address in handle */
    if (recpaddr)
      (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
          (void *) recpaddr, (ub4) strlen(recpaddr),
          (ub4) OCI_ATTR_SUBSCR_RECPT, errhp);

    /* set receive with presentation in handle */
    (void) OCIAttrSet((void *) *subscrhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
        (void *) &recppres, (ub4) 0,
        (ub4) OCI_ATTR_SUBSCR_RECPTPRES, errhp);

    printf("Begining Registration for subscription %s\n", subscriptionName);
    checkerr(errhp, OCISubscriptionRegister(svchp, subscrhp, 1, errhp,
        OCI_DEFAULT));
   printf("done\n");

}


int main( argc, argv)
int    argc;
char * argv[];
{
    OCISession *authp = (OCISession *) 0;

/*****************************************************
Initialize OCI Process/Environment
Initialize Server Contexts
Connect to Server
Set Service Context
******************************************************/

/* Registration Code Begins */
/* Each call to initSubscriptionHn allocates
   and Initialises a Registration Handle */

/* Register for OCI notification */
    initSubscriptionHn(    &subscrhpSnoop,    /* subscription handle*/
    (char*) "PUBSUB.LOGON:SNOOP", /* subscription name */
                                  /*<queue_name>:<agent_name> */
        (void*)notifySnoop,  /* callback function */
        OCI_SUBSCR_PROTO_OCI, /* receive with protocol */
        (char *)0, /* recipient address */
        OCI_SUBSCR_PRES_DEFAULT); /* receive with presentation */

/* Register for email notification */
    initSubscriptionHn(    &subscrhpSnoopMail,  /* subscription handle */
     (char*) "PUBSUB.LOGON:SNOOP",              /* subscription name */ 
                                                /* <queue_name>:<agent_name> */
        (void*)0, /* callback function */
        OCI_SUBSCR_PROTO_MAIL, /* receive with protocol */
        (char*)  "longying.zhao@oracle.com", /* recipient address */
        OCI_SUBSCR_PRES_DEFAULT); /* receive with presentation */

/* Register for server to server notification */
    initSubscriptionHn(    &subscrhpSnoopServer, /* subscription handle */
       (char*)  "PUBSUB.LOGON:SNOOP",            /* subscription name */
                                                 /* <queue_name>:<agent_name> */
        (void*)0, /* callback function */
        OCI_SUBSCR_PROTO_SERVER, /* receive with protocol */
         (char*) "pubsub.plsqlnotifySnoop", /* recipient address */
        OCI_SUBSCR_PRES_DEFAULT); /* receive with presentation */

    checkerr(errhp, OCITransCommit(svchp, errhp, (ub4) OCI_DEFAULT));

/*****************************************************
The Client Process does not need a live Session for Callbacks
End Session and Detach from Server
******************************************************/

    OCISessionEnd ( svchp,  errhp, authp, (ub4) OCI_DEFAULT);

    /* detach from server */
    OCIServerDetach( srvhp, errhp, OCI_DEFAULT);

   while (1)    /* wait for callback */
    sleep(1);
}

If user IX logs on to the database, the client is notified by e-mail, and the callback function notifySnoop is called. An e-mail notification will be sent to the address xyz@company.com and the PL/SQL procedure plsqlnotifySnoop will also be called in the database.

Publish-Subscribe LDAP Registration Example

The following code fragment illustrates how to do LDAP registration. Please read all the program comments:

...

  /* TO use LDAP registration feature, OCI_EVENTS | OCI_EVENTS |OCI_USE_LDAP must
  /*   be set in OCIInitialize: */
  (void) OCIInitialize((ub4) OCI_EVENTS|OCI_OBJECT|OCI_USE_LDAP, (void *)0,
                       (void * (*)(void *, size_t)) 0,
                       (void * (*)(void *, void *, size_t))0,
                       (void (*)(void *, void *)) 0 );

...

  /* set LDAP attributes in the environment handle */

  /* LDAP host name */
  (void) OCIAttrSet((void *)envhp, OCI_HTYPE_ENV, (void *)"yow", 3,
                    OCI_ATTR_LDAP_HOST, (OCIError *)errhp);

  /* LDAP server port */ 
  ldap_port = 389;
  (void) OCIAttrSet((void *)envhp, OCI_HTYPE_ENV, (void *)&ldap_port,
                    (ub4)0, OCI_ATTR_LDAP_PORT, (OCIError *)errhp);

  /* bind DN of the client, normally the enterprise user name */
  (void) OCIAttrSet((void *)envhp, OCI_HTYPE_ENV, (void *)"cn=orcladmin",
                    12, OCI_ATTR_BIND_DN, (OCIError *)errhp);

  /* password of the client */
  (void) OCIAttrSet((void *)envhp, OCI_HTYPE_ENV, (void *)"welcome",
                    7, OCI_ATTR_LDAP_CRED, (OCIError *)errhp);

  /* authentication method is "simple", username/password authentication */
  ldap_auth = 0x01;
  (void) OCIAttrSet((void *)envhp, OCI_HTYPE_ENV, (void *)&ldap_auth,
                    (ub4)0, OCI_ATTR_LDAP_AUTH, (OCIError *)errhp);

  /* adminstrative context: this is the DN above cn=oraclecontext */
  (void) OCIAttrSet((void *)envhp, OCI_HTYPE_ENV, (void *)"cn=acme,cn=com",
                    14, OCI_ATTR_LDAP_CTX, (OCIError *)errhp);

...

  /* retrieve the LDAP attributes from the environment handle */

  /* LDAP host */
  (void) OCIAttrGet((void *)envhp, OCI_HTYPE_ENV, (void *)&buf, 
                    &szp,  OCI_ATTR_LDAP_HOST,  (OCIError *)errhp);

  /* LDAP server port */
  (void) OCIAttrGet((void *)envhp, OCI_HTYPE_ENV, (void *)&intval, 
                    0,  OCI_ATTR_LDAP_PORT,  (OCIError *)errhp);

  /* client binding DN */
  (void) OCIAttrGet((void *)envhp, OCI_HTYPE_ENV, (void *)&buf, 
                    &szp,  OCI_ATTR_BIND_DN,  (OCIError *)errhp);

  /* client password */
  (void) OCIAttrGet((void *)envhp, OCI_HTYPE_ENV, (void *)&buf, 
                    &szp,  OCI_ATTR_LDAP_CRED,  (OCIError *)errhp);

  /* adminstrative context */
  (void) OCIAttrGet((void *)envhp, OCI_HTYPE_ENV, (void *)&buf, 
                    &szp,  OCI_ATTR_LDAP_CTX,  (OCIError *)errhp);

  /* client authentication method */
  (void) OCIAttrGet((void *)envhp, OCI_HTYPE_ENV, (void *)&intval, 
                    0,  OCI_ATTR_LDAP_AUTH,  (OCIError *)errhp);
  
  ...

  /* to set up the server DN descriptor in the subscription handle */

  /* allocate a server DN descriptor, dn is of type "OCIServerDNs **", 
     subhp is of type "OCISubscription **" */
  (void) OCIDescriptorAlloc((void *)envhp, (void **)dn, 
                         (ub4) OCI_DTYPE_SRVDN, (size_t)0, (void **)0);

  /* now *dn is the server DN descriptor, add the DN of the first database 
     that we want to register */
  (void) OCIAttrSet((void *)*dn, (ub4) OCI_DTYPE_SRVDN, 
                    (void *)"cn=server1,cn=oraclecontext,cn=acme,cn=com",
                    42, (ub4)OCI_ATTR_SERVER_DN, errhp);
  /* add the DN of another database in the descriptor */
  (void) OCIAttrSet((void *)*dn, (ub4) OCI_DTYPE_SRVDN, 
                    (void *)"cn=server2,cn=oraclecontext,cn=acme,cn=com",
                    42, (ub4)OCI_ATTR_SERVER_DN, errhp);

  /* set the server DN descriptor into the subscription handle */
  (void) OCIAttrSet((void *) *subhp, (ub4) OCI_HTYPE_SUBSCRIPTION,
                 (void *) *dn, (ub4)0, (ub4) OCI_ATTR_SERVER_DNS, errhp);

  ...

  /* now we will try to get the server DN information from the subscription
     handle */
 
  /* first, get the server DN descriptor out */
  (void) OCIAttrGet((void *) *subhp, (ub4) OCI_HTYPE_SUBSCRIPTION, 
                    (void *)dn, &szp, OCI_ATTR_SERVER_DNS, errhp);

  /* then, get the number of server DNs in the descriptor */
  (void) OCIAttrGet((void *) *dn, (ub4)OCI_DTYPE_SRVDN, (void *)&intval,
                    &szp, (ub4)OCI_ATTR_DN_COUNT, errhp);

  /* allocate an array of char * to hold server DN pointers returned by
     oracle */
    if (intval)
    {
      arr = (char **)malloc(intval*sizeof(char *));
      (void) OCIAttrGet((void *)*dn, (ub4)OCI_DTYPE_SRVDN, (void *)arr,
                        &intval, (ub4)OCI_ATTR_SERVER_DN, errhp);
    }

  /* OCISubscriptionRegister() calls have two modes: OCI_DEFAULT and 
     OCI_REG_LDAPONLY. If OCI_DEFAULT is used, there should be only one
     server DN in the server DN descriptor. The registration request will
     be sent to the database. If a database connection is not available,
     the registration request will be detoured to the LDAP server. On the 
     other hand, if mode OCI_REG_LDAPONLY is used the registration request
     will be directly sent to LDAP. This mode should be used when there are 
     more than one server DNs in the server DN descriptor, or we are sure
     that a database connection is not available.

     In this example, two DNs are entered; so we should use mode 
     OCI_REG_LDAPONLY in register. */
  OCISubscriptionRegister(svchp, subhp, 1, errhp, OCI_REG_LDAPONLY);

  ...

  /* as OCISubscriptionRegister(), OCISubscriptionUnregister() also has
     mode OCI_DEFAULT and OCI_REG_LDAPONLY. The usage is the same. */

  OCISubscriptionUnRegister(svchp, *subhp, errhp, OCI_REG_LDAPONLY);
}
...