Thursday 4 June 2020

How to use coe_xfr_sql_profile.sql to Improve performance of a sql query?

Oracle SQL Performance tuning is one of the frequently reported issues in a Oracle Database administrator's daily job.
Very often we notice that sql queries or concurrent Programs which were running within few minutes earlier are now taking hours to be completed.
Underlying reason is due to the plan_hash_value used by the query has been changed.
coe_xfr_sql_profile.sql is the script developed by Oracle Support which helps us to identify the plan_hash_value which will resolve the performance issue.

Steps to use the script coe_xfr_sql_profile.sql

1. Download the script to a temporary directory on the database server.
Script is also available in Oracle Metalink.

2. Before Running the coe_xfr_sql_profile.sql, we need to know the sql_id of the sql query which is taking longer time to complete than usual.

To find the sql_id use below query

2.1 If you know the database SID of the long running query.
select inst_id,sid,serial#,status,last_Call_et,sql_id from gv$session where sid= '&sessionid';

2.2 If you know the query test, Use gv$sql and gv$sqlarea views to get the sql_id.

3. Login to sqlplus as sysdba and run coe_xfr_sql_profile.sql at the SQL prompt.

$ sqlplus “/as sysdba”
SQL> @coe_xfr_sql_profile.sql
Parameter 1:
SQL_ID (required)

Enter value for 1: 0a3f7vuks8d7y   (--this is the sql_id of long running sql)
PLAN_HASH_VALUE AVG_ET_SECS
————— ———–
1484137450 15.6
3622468234 4560.76

[Output shows the list of available plan hash values, which can be forced on to the sql query. We need to select the plan_hash_value with low ETA to resolve the performance issue]

Parameter 2:
PLAN_HASH_VALUE (required)

Enter value for 2: 1484137450 


++Based on the inputs provided, it generates a sql script with naming convention coe_xfr_sql_profile_0a3f7vuks8d7y_1484137450.sql as output

coe_xfr_sql_profile_<sql_id>_<plan_hash_value>.sql

4. Run the script as sysdba

SQL>coe_xfr_sql_profile_0a3f7vuks8d7y_1484137450.sql 

5. Verify that Performance issue is resolved by Re-running the sql.

No comments:

Post a Comment

Was this Post Helpful?

Feel free to suggest your opinions in the comments section!