Reg : -fix_control parameter

[Former Member]

Hi All,

I was going through the note 1165319 regarding the patch 7188932 is contained in the Optimizer merge fix #3 (7228971).

The below statement was given in the note.Could anyone help me in understanding as to what it means.When i check DB01 in my systems i cannot see anything.The below explanation is going over my head.Please dont consider that i am lazy and have posted it.I had done my research and as a last resort posted the question.Thanks.

An SQL query does not complete or the process terminates with a timeout.

In DB01 or in V$LOCK, the system displays a "CU" lock for the process. V$SESSION_WAIT displays the event "db file sequential read" over a long time. However, the sequence number (SEQ#) does not change.

The parameters P1 and P2 in V$SESSION_WAIT display Block=<n> for FILE=1. Here, the block belongs to the Oracle Cluster C_OBJ#_INTCOL#. An Oracle error stack shows that the process in the function kkompr() loops.

This system response occurs during the parse for the statement. Therefore, you can use EXPLAIN PLAN FOR Select ... to reproduce it.

[Former Member · Accepted solution]

Hi, Ambarish

The note says that you may face with CU (cursor sharing) lock, during the process. That may be occured sometimes, regarding your database requests. In order to watch this problem, you can execute "EXPLAIN PLAN FOR Select ..." statement. So, the system will generate a CU lock and show it in the V$LOCK table. You can also monitor the problem by checking V$SESSION_WAIT table. During the process, you will see that the database is reading same data block on Oracle which is problem itself.

As a summary of it, Oracle kills the database request with the "CU lock", because of timeout. In order to avoid this problem, you need to apply this patch on the Oracle.

I faced with this problem, previously and solved it with this patch.

Best regards,

Orkun Gedik