1. Installation

 Installing the statspack is an easy step. The only two things that need to be done is to execute the statspack creation script that is located in %OracleHome%\rdbms\admin folder. The script is named as “spcreate.sql” for creating the statspack objects and “spdrop.sql” to drop the statspack user and its associated objects. The second thing though not necessary that you can do is to dedicate a seperate tablespace for statspack, i’ve created a seperate tablespace named STATSPACKTBS for it.

In oder to install the statspack, use the following comand replacing the location where the script is located:

SQL> @c:\oracle\rdbms\admin\spcreate.sql;

During the execution of the script, it will ask three things from the user (1) the password for the username of statspack which is “perfstat”.  (2) Default tablespace for statspack (mine is dedicated STATSPACKTBS), and (3) Temporary tablespace for use.

2.  Generating Snapshots

In order to generate statspack report, first a snapshot of the database needs to be taken so that the analysis can be done. The report is always generated between any two snapshots. To generate a snapshot use the following command:

SQL> execute statspack.snap;

3.  Generating Statspack Report

In order to generate statspack report, we only have to execute another script supplied by oracle located in rdbms\admin folder named as “spreport.sql”. To generate a report, it is necessary that you must have atleast two snapshots of the database otherwise lots of portions of the report would not be analysed or calculated by the statspack tool. To generate a report, use the following command :

SQL> @C:\oracle\rdbms\admin\spreport.sql;

At the start of the report, it will show the list of snapshots currently present and would onwards ask for the beginning and ending snapshot ID to compare and analyse.

Although executing sql statements such as select statement and performing DML statements are easy using oracle HS database link. This is not the same case with executing functions and procedures of the remote DB from oracle environment. For executing a procedure, Oracle has provided a DBMS package known as DBMS_HS_Passthrough which actually sends the request to the remote db for execution.

             In order to execute the procedure at the remote end, we need to utilize the EXECUTE_IMMEDIATE procedure to send the remote db a request to execute the command included as a parameter. For example, i want to execute a remote db procedure ‘receivedoc’. The remote procedure ‘receivedoc’ takes a single input value param1 as varchar and has no return value. So to execute this from oracle env. i will have to execute the following command

               DBMS_HS_Passthrough.Execute_Immediate(‘execute receivedoc ”value1”;’);

            One thing to keep in mind is that the DBMS_HS_PASSTHROUGH package returns a value for success/failure. Therefore executing the above statement at the sql prompt will give error. Therefore, we need to wrap it around a pl/sql block to capture the return value as below:

declare
        ret number;
begin
        ret := DBMS_HS_Passthrough.Execute_Immediate(‘execute receivedoc ”value1”;’);
end;

           Second thing to keep in mind is that oracle requires a commit after a DML operation completes at either end. Therefore, if a remote procedure has performed DML statements, commit is required to be issued after it to release any locks applied by either side. One problem that i encountered with this practice was the table locks after the procedure completes. This was because the procedure issued DML statements over it and didn’t perform commit to release the locks (a simple thing ) … Issuing a commit on either side releases the locks, however, it is recommended that the commit be issued within the remote db procedure rather than than oracle pl/sql block. This is simply to avoid intruption due to any connection breakage e.g network, and ultimately leaving the table in locks permanently.

Hope this post may serve you ppl with some knowledge…

Cheers !

1. Shutdown the database using
sql> shutdown normal

2. Add only one entry parameter namely ‘IFILE’ in a newly created pfile as below

IFILE = ‘F:\BACKUP LOCATION\spfileORCL.ORA’

3. Save the file at the default or any desired location e.g. C:\pfileORCL.ora

4. Start the database using the following command while specifying the pfile location….

sql> startup pfile=’C:\pfileORCL.ora’

5. After database startup, verify whether the instance has been started up with spfile using the below command:
sql> sho parameter spfile

If the value column contains a non null value for the spfile parameter, it would indicate that the instance has used spfile for startup, which in our case is at the non-default location. That is the only option available in oracle 9i to startup the oracle instance from a non-default location, no other option exists.

Thanks…

Happy learning …

Cheers !….

Oracle Heterogeneous services provide a generic method for connecting to any non-oracle ODBC or OLEDB compliant database such as MS Acess, SQL Server etc. One of the other services by oracle is known as Oracle Transparent Gateways which in contrast to Oracle Heterogeneous services provide a more optimized and better functionality for connecting to non-oracle systems. For simple connectivity, I utilized the heterogeneous service that is also simple and easy to configure.

Creating a System DSN

In order to connect to sql server database, we need a system dsn which is used by the oracle services for connecting to the remote sql server. A brief guide on how to create a system dsn for sql server is available at http://www.truthsolutions.com/sql/odbc/creating_a_new_odbc_dsn.htm

Setting up heterogeneous services configuration file

For telling the oracle about the DSN, create an HS agent file named as init.ora in your HS\Admin subdirectory. That is, if you configured your DSN name as sqldb, you would need to create file as initSQLDB.ora in your hs\admin sub directory under your oracle home. Open the file and enter two parameters in the agent file as below:

• HS_FDS_CONNECT_INFO = SQLDB
• HS_FDS_TRACE_LEVEL = OFF

Configuring tnsnames.ora

After the HS agent file has been setup, configure the oracle TNS listener for an entry of the odbc dsn so that it can utilize the dsn. Enter the parameter lines as below:

(CONNECT_DATA = (SID = SQLDB))
(HS=OK)

where, HS=OK indicates the oracle server to utilize heterogeneous service named (SID = SQLDB) which is eventually the initSQLDB.ora file previously created and configured. A configured sample file is as below:-

sqldb.world =
(DESCRIPTION =
(ADDRESS_LIST =
(ADDRESS = (PROTOCOL = TCP)(HOST = localhost)(PORT = 1521))
)
(CONNECT_DATA =(SID = sqldb))
(HS = OK)
)

Configuring listener.ora

After configuring tnsnames.ora file, we need to add the sqldb dsn program type in the listener.ora file as below:-

(SID_DESC =
(SID_NAME = sqldb)
(ORACLE_HOME = d:\oracle\ora92)
(PROGRAM = hsodbc)

This would provide the listener information about the type of connection to use i.e. hsodbc or heterogeneous service in this case. After the listener.ora and tnsnames.ora files have been configured, we need to reload the new setting in to the listener to take effect. For this, reload or restart the tns listener by issuing below mentioned commands:

C:\> lsnrctl reload

OR

C:\> lsnrctl
lsnrctl> stop
lsnrctl> start

Creating a Database link

Now that we have setup the configuration, final thing to do is to create a database link to connect to sql server database. For this, we must have a valid sql server username and password setup with all the required roles and privileges for accessing the sql server database. Use the following command to create a database link:

SQL> create public database link MySQLConn connect to “MyUser” identified by “MyPass” using ‘SQLDB.WORLD’;

Where,
MyUser / MyPass indicate a valid username and password of SQL Server database.

Querying sql server database

By utilizing the above database link, we can access the sql server database tables. e.g. below sql statement:

1. Provides a method of backing and archiving historical data.
2. Provides a way for saving a baseline structure of table definitions that can be used in case of user errors
3. Provides export and import between different versions of Oracle server.
4. Provides method of transporting tablespaces between different databases.

Methods of Utility Invocation

Both the export and import utilities can be invoked through two different interfaces namely (1) command line (2) using Enterprise Manager (OEM). The command line interface however provides more flexibility in terms of parameters than OEM. The command line interface can be used with direct command line parameters or by providing a parameter file to export/import utility. The command line interface also provides an interactive dialog for completing the tasks. In order to use the export / import utility, the user must have ‘CREATE SESSION’ and ‘EXP_FULL_DATABASE’ system privileges.

Export / Import Modes

There are four modes of exporting or importing data. Each mode provides a seperate level of degree for exporting and importing data from the database. There four modes are :

1. Table Mode provides facility to the user in exporting or importing a specific table rather than exporting / importing all tables.

2. User Mode provide facility in exporting / importing all objects of a specified user. This includes all table definitions, table data, indexes, owner grants and related constraints.

3. Tablespace Mode provides facility in exporting / importing all objects of a given tablespace. The transportable tablespace feature enables to move a set of tablespaces from one oracle database to another even if the block sizes are different. Transportable tablespace feature can be much faster than a normal export / import as it only integrates tablespace structural information with the datafile copies.

4. Full Database Mode provides facility in exporting all database objects except the objects of SYS schema.

Executing Export / Import Utility using Command line Interface

Below is a simple example of export and import command to export/import a user HR.

C:\> exp system/manager OWNER=HR direct=Y file=c:\dumps\export-HR.dmp

where,
               ‘OWNER’ indicates a user-mode export and lists the users whose objects are required to be exported.
               ‘DIRECT’ indicates the use of direct path method that provides much faster export bypassing the sql command processing layer and reading directly from the data layer.
              ‘FILE’ indicates the location and name of the binary export file.

C:\> imp system/manager FROMUSER=HR   TOUSER=DB2HR   file=c:\export-HR.dmp

where,
               ‘FROMUSER’ indicates the user from which the data is extracted by import utility from the export file
               ‘TOUSER’ indicates the user to which the data is to be imported.
               ‘FILE’ indicates the location and name of the binary export file.

Any Suggessions and comments are welcome … !!!!!

Thanks…!!