Skip navigation.

Vivek Sharma

Syndicate content
I have consolidated, and will do so in future, all my Experiences for Oracle Community. All these are my own opinions and does not necessarily reflect the views of Oracle. There are many databases with different permutation and combination which I have not seen. Hence, with these limitations, my views might be wrong, before implementing do test these. Please feel free to comment on my views, be it a Criticism, as this will only improve my writing abilities.Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.comBlogger55125
Updated: 8 min 42 sec ago

Performance Issues after Upgrading to 10g Release 2 - Apps Upgrade 11.5.9 to 11.5.10

Mon, 2009-01-05 12:27
Upgrading from 9i to 10g has been giving sleepness nights to many of the dba's and application team. There are many changes that has been incorporated in 10g, like, method_opt that defaults to "for all columns size auto", System Statistics is mandatory in 10g etc. Optimizer behaviour has also significantly changed between the two versions, therefore, a proper testing of the entire application Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com1

Database Freeze when SGA_TARGET set to a high value

Mon, 2008-12-15 23:48
During one of my technical presentation organized by "All India Oracle User Group", I mentioned a point under "Beaware of :" section, wherein, I mentioned something on Oracle New Features. With each new version, Oracle introduced new features, that certainly help eliminate one contention, but might introduce new issues and this new issue is sometimes difficult to relate to the newly introduced Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com4

Database Freeze when SGA_TARGET set to a high value

Mon, 2008-12-15 23:48
During one of my technical presentation organized by "All India Oracle User Group", I mentioned a point under "Beaware of :" section, wherein, I mentioned something on Oracle New Features. With each new version, Oracle introduced new features, that certainly help eliminate one contention, but might introduce new issues and this new issue is sometimes difficult to relate to the newly introduced Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com2

Literals v/s Bind Variables as an Input / Output parameter to a Procedure

Tue, 2008-11-18 11:58
My previous blog was on Library Cache Latch Contention and in this blog too, I am going to discuss something that has an Impact on Latches, especially Library Cache Latches. During my investigation on a Severe Library Cache Latch Contention, I came across many procedures that are executed through out the day and surprisingly use Literals as an Input Parameters. There were around 30 to 50 such Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com2

Literals v/s Bind Variables as an Input / Output parameter to a Procedure

Tue, 2008-11-18 11:58
My previous blog was on Library Cache Latch Contention and in this blog too, I am going to discuss something that has an Impact on Latches, especially Library Cache Latches. During my investigation on a Severe Library Cache Latch Contention, I came across many procedures that are executed through out the day and surprisingly use Literals as an Input Parameters. There were around 30 to 50 such Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com2

Library Cache Latch Contention and Scalability

Sun, 2008-11-09 20:51
Lot has been written about Library Cache Latch Contention and one of the primary reason for this being "Too many Hard / Soft Parses". I have couple of blogs on this issue that was due to improper OS settings. There are many White papers that mention about this contention and all of these primarily state about LC Latches due to Hard or Soft Parses. To get a good conceptual knowledge on this Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com5

Library Cache Latch Contention and Scalability

Sun, 2008-11-09 20:51
Lot has been written about Library Cache Latch Contention and one of the primary reason for this being "Too many Hard / Soft Parses". I have couple of blogs on this issue that was due to improper OS settings. There are many White papers that mention about this contention and all of these primarily state about LC Latches due to Hard or Soft Parses. To get a good conceptual knowledge on this Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com4

Optimizer Behaviour in 10g R2 - Importance of Fresh Optimizer Statistics

Sat, 2008-10-25 01:52
On 19th October I wrote about a discussion with a customer on the importance on Statistics after Merging of Data. The scenario that I mentioned was that the customer planned to merge data from database C to H and to differentiate this data, some of the critical (queried) columns of C will be incremented by 900000000. Since, many of these columns are Unique / Primary keys, there ar possibilities Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com1

Optimizer Behaviour in 10g R2 - Importance of Fresh Optimizer Statistics

Sat, 2008-10-25 01:52
On 19th October I wrote about a discussion with a customer on the importance on Statistics after Merging of Data. The scenario that I mentioned was that the customer planned to merge data from database C to H and to differentiate this data, some of the critical (queried) columns of C will be incremented by 900000000. Since, many of these columns are Unique / Primary keys, there ar possibilities Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com1

Optimizer behaviour in 10g Release 2 - A real life scenario

Sat, 2008-10-18 22:22
Recently I was involved in a technical discussion on the necessity to gather fresh statistics after merging data from a database into another. The background of this activity is as under : 1. Data from C Database will be merged into H (I do not want to expose the customer name and the db name, therefore, will use these names). 2. The current size of H database is 800 GB and after the merger, it Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com1

Optimizer behaviour in 10g Release 2 - A real life scenario

Sat, 2008-10-18 22:22
Recently I was involved in a technical discussion on the necessity to gather fresh statistics after merging data from a database into another. The background of this activity is as under : 1. Data from C Database will be merged into H (I do not want to expose the customer name and the db name, therefore, will use these names). 2. The current size of H database is 800 GB and after the merger, it Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com1

Oracle Cost Based Optimizer on AIOUG

Wed, 2008-10-08 02:01
I will be presenting a 1 hour session on "Oracle Cost Based Optimizer". This is a second technight organized by All India Oracle User Group and is scheduled to be held at Bangalore on 22nd October 2008.Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com0

Oracle Cost Based Optimizer on AIOUG

Wed, 2008-10-08 02:01
I will be presenting a 1 hour session on "Oracle Cost Based Optimizer". This is a second technight organized by All India Oracle User Group and is scheduled to be held at Bangalore on 22nd October 2008.Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com0
Categories: Fusion Middleware

Multiple Version Count - BIND_PEEKED_PQ_MISMATCH & PQ_SLAVE_MISMATCH and few more....

Sat, 2008-09-20 10:45
Recently, during a generic health check of a critical database at a customer site, I noticed high values of Version Count for 3-4 Queries. I have a blog on this topic, in the month of June 2008 and this is something new that I noticed and thought this to be worth writing. The database version in this case is 10g Release 2. In this case, there were three type of issues that were contributing to Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com0

Multiple Version Count - BIND_PEEKED_PQ_MISMATCH & PQ_SLAVE_MISMATCH and few more....

Sat, 2008-09-20 10:45
Recently, during a generic health check of a critical database at a customer site, I noticed high values of Version Count for 3-4 Queries. I have a blog on this topic, in the month of June 2008 and this is something new that I noticed and thought this to be worth writing. The database version in this case is 10g Release 2. In this case, there were three type of issues that were contributing to Vivek Sharmahttp://www.blogger.com/profile/08767102375351411984vivek.s.sharma@oracle.com0