Archive

Archive for the ‘Database’ Category

Gather Schema and Index Statistics in Oracle

January 28th, 2010 2 comments

In this post I would like to draw your attention to the Optimizer in Oracle 10g.
The best place to start is the DBMS_STATS package reference page.

The DBMS_STATS package contains all the procedures required to maintain the optimizer statistics of any schema.
Once you get familiar with the documentation of the package, you will actually know which approach is best for your database and schemas. In my case I have to gather schema statistics once per week with a specific estimate_percent value and I have to gather schema index statistics every night. For this purpose I have created the following simple PL/SQL. I recommend if you create jobs and schedule them as appropriate.

-- Gathering Schema Statistics
execute dbms_stats.gather_schema_stats('SCHEMA1', estimate_percent=>39, cascade=>TRUE);
execute dbms_stats.gather_schema_stats('SCHEMA2', estimate_percent=>45, cascade=>TRUE);
-- Gathering Index Statistics
set serveroutput on
begin
for ind in
(select object_name
from user_objects
where object_type='INDEX'
and object_name like 'YOUR_INDEX_PATTERN_HERE'
)
loop
dbms_output.put_line('Gathering Index Statistics for '||ind.object_name||'.....');
dbms_stats.gather_index_stats('SHEMA1', ind.object_name , estimate_percent=>100);
dbms_output.put_line('Gathering Index Statistics for '||ind.object_name||' is Complete!');
end loop;
end;
/

Read more…

Compile Invalid SYNONYM Oracle

January 11th, 2010 4 comments

Happy New Year!
Wish you all the great stuff you have been dreaming about to come true in 2010!

I haven’t posted for a long time but my excuse is the busy season at work. At the end of the day I am finding a minute to share a simple but useful hint on generating the COMPILE SQL for invalid PUBLIC SYNONYM objects.

Every time you drop and recreate an object such as a SEQUENCE, the PUBLIC SYNONYM for that objects invalidates (if such exists). In order to generate the COMPILE commands you could use SQL similar to the following two examples:

Read more…

Categories: Database, Oracle Tags: , , , ,

Fixing incorrect crs_stat

September 8th, 2009 Comments off

I was interested to find out how I can change the configuration of the oracle cluster nodeapps.
I had a problem with the listener running on both nodes.
The problem was an incorrect ACTION_SCRIPT=/u01/app/oracle/oracle/product/10.2.0/asm/bin/racgwrap
The path was incorrect due to previous installation that I did with the wrong path. This leads to unknown status in the application state as follows:

oracle@rac03:/u01/crs1020/crs/public> crs_stat -t
Name Type Target State Host
————————————————————
ora….SM1.asm application ONLINE ONLINE rac03
ora….03.lsnr application ONLINE UNKNOWN rac03
ora.rac03.gsd application ONLINE ONLINE rac03
ora.rac03.ons application ONLINE ONLINE rac03
ora.rac03.vip application ONLINE ONLINE rac03
ora….SM2.asm application ONLINE ONLINE rac04
ora….04.lsnr application ONLINE UNKNOWN rac04
ora.rac04.gsd application ONLINE ONLINE rac04
ora.rac04.ons application ONLINE ONLINE rac04
ora.rac04.vip application ONLINE ONLINE rac04

In order go fix this I exported the listener configuration by using the following for both nodes separately:

Read more…

Invalid Objects after wrong utlrp.sql

June 24th, 2009 Comments off

Be extra careful with the location of your utlrp.sql script when you decide to recompile the invalid objects for the entire database as sys.
The utlrp.sql script resides in the ORACLE_HOME/rdbms/admin and in CLIENT_HOME/rdbms/admin folders.
It is utterly wrong to use and recompile with the script in the client home folder.
After such a mistake of me the SYSMAN schema got completely corrupt.
I restored the SYSMAN using the steps described in my previous post OracleDBControl .

The remaining invalid objects are as follows:

select owner, object_name from all_objects where status='INVALID';

Read more...

Categories: Oracle Tags:

OracleDBControl

June 24th, 2009 Comments off

If you have ever exported and imported complete databases and moved them around from one server to another, you have probably noticed the issues that might occur with the EM repository. During the last delivery project that I worked on I had to import two databases in two Oracle servers (10.2.0.4) installed on Microsoft Windows 2003 x64. Everything was going smooth with the DBCA during the creation of the new database. I have selected the database to be manager using Enterprise Manager so the EM repository was being installed. Unfortunately it errored out. The logs are stored in the following location:

D:\oracle\product\10.2.0\db_1\cfgtoollogs\emca\instance_name (ORACLE_HOME\cfgtoollogs\emca\instance_name)

I found out that everything with the creation of the EM repository has been successful except the actual start of the windows service. In the log files I found the following trace:

24.06.2009 10:11:50 oracle.sysman.emcp.EMReposConfig createRepository
INFO: Creating the EM repository (this may take a while) ...
24.06.2009 10:14:32 oracle.sysman.emcp.EMReposConfig invoke
INFO: Repository successfully created
24.06.2009 10:14:42 oracle.sysman.emcp.util.DBControlUtil secureDBConsole
INFO: Securing Database Control (this may take a while) ...
24.06.2009 10:15:08 oracle.sysman.emcp.util.DBControlUtil secureDBConsole
INFO: Database Control secured successfully.
24.06.2009 10:19:48 oracle.sysman.emcp.util.DBControlUtil startOMS
INFO: Starting Database Control (this may take a while) ...
24.06.2009 10:21:13 oracle.sysman.emcp.util.DBControlUtil startOMS
INFO: Starting Database Control (this may take a while) ...
24.06.2009 10:25:08 oracle.sysman.emcp.util.DBControlUtil startOMS
INFO: Starting Database Control (this may take a while) ...
24.06.2009 10:24:08 oracle.sysman.emcp.util.DBControlUtil startOMS
INFO: Starting Database Control (this may take a while) ...

Read more…

Categories: Oracle Tags:

Drop multiple tables in Oracle

April 23rd, 2009 13 comments

I just noticed that the Oracle instance that I am working with has the SMP_% tables added from the Oracle Enterprise Manager. Since I don’t need them I decided to drop them. Here is my sql that generates the delete statements:

select ‘drop table ‘||table_name||’;’ from user_tables where table_name like ‘SMP_%’;

or..

Read more…

Categories: Oracle Tags: ,

Extending VMDK file and resizing the partition – SLES 10 Linux

April 7th, 2009 Comments off

1. ESXi 3.5 Extending the VMDK file [Virtual Machine Hard Disk]

The main reason behind this is the fact that I needed more space for 2 more Oracle databases on a SLES10 Linux system. The partition mounted on /u01 was initially created as 21GB but I quickly depleted the space with 3 oracle 10g databases that took more than 17GB and the space left was not sufficient for the 2 new databases that I had to create.

Before I advanced with the extension of the VMDK file I connected to each separate instance and performed “shutdown immediate” command as sysdba.
Then I stopped the listener, dbconsole, isqlplus and once I confirmed that no oracle related processes were present on the system I performed a shutdown “shutdown -h now”.

In order to extend an existing hard drive attached to a virtual machine you have to make sure no snapshots of the virtual machine are present. I know that this is quite uncomfortable considering the risky operation that you are about to perform but there is a work around. (The work around is not in this post, so please let me know if you are interested or simply search for it. There is a good chance that I will have an article concerning that topic)

Read more…

Oracle 10g RAC on ESXi3 using SLES9 SP5 – Part 10

February 20th, 2009 3 comments

When I finished the clusterware install in the lab environment I have continued with the database software itself but it appears that I needed a VMware ESXi Update 3 required! The kernel version of SLES failed verification and upon an upgrade of the kernel through YAST both machines were unusable since they would not boot.

This is an obstacle until completed since I can’t continue with the How-to until this is done.

In Part 9 of the Oracle RAC How-to we have completed successfully the installation of the Oracle Clusterware services on the shared storage for the two Suse Linux Enterprise Server 9 SP5 nodes.

Read more…

Oracle alter table change column datatype

February 17th, 2009 Comments off

Let’s say you had a field CITIZEN in table called EMPLOYEE that was created as VARCHAR2(10) but this has to be changed to VARCHAR2(3). The syntax to change the the column datatype in Oracle is as follows:


ALTER TABLE table_name MODIFY ( column_name VARCHAR2(3));


If there is existing column data you should first export that data and then import it after the datatype change. Make sure the exported data will be suitable for import upon the change (You can’t import a row with CITIZEN column of 5 after you have changed the column to VARCHAR2(3)).

Categories: Database, Oracle Tags: , ,

Oracle 10g RAC on ESXi3 using SLES9 SP5 – Part 9

February 16th, 2009 Comments off

Verify the Oralce Clusterware Installation

With the introduction of Oracle RAC 10g, cluster management is controller by the evmd, ocssd and crsd processes.
Run the ps command on both nodes to make sure that the processes are running.


rac01:/u01/clusterware/cluvfy # ps -ef |grep d.bin
root 4694 1 0 Feb13 ? 00:00:00 /u01/crs1020/bin/crsd.bin reboot
oracle 5242 4692 0 Feb13 ? 00:00:00 /u01/crs1020/bin/evmd.bin
oracle 5344 5326 0 Feb13 ? 00:00:00 /u01/crs1020/bin/ocssd.bin
root 20078 10946 0 09:44 pts/1 00:00:00 grep d.bin

Next you should check the /etc/inittab file, which is processed whenever the runlevel changes:

Read more…