BIAS Recommended Settings
BIAS Recommended Settings
DB initialization parameter
BIAS Recommended settings
For Release 12, a reasonable starting point is a size of 600M and a 60M reserved area (10%).
shared_pool_size = 600M
Comment:-As per note 396009.1 we will set up the initialization parameters as per the number of users. We need
the number of users to set this parameter in Prod and Pre-Prod
Also, we are using the automatic shared memory management feature by setting the sga_target. So, the
shared_pool, buffer cache and redo log buffer will be automatically tuned based on the requirement. A value of
400M is a starting point and automatic SGA tuning will adjust the caches as per the workload.
2. DB initialization parameter
Recommendations
PL/SQL native compilation is recommended for Oracle Database 11g based Oracle E-Business Suite environments
such as Release 12 to maximize runtime performance and scalability.
*.plsql_code_type='NATIVE'
The below parameters can be removed from the initialization parameter file since this is a 11gRel1 DB.
*.plsql_native_library_dir='/backup/SFSXS5Q/db/tech_st/11.1.0/plsql/nativelib'
*.plsql_native_library_subdir_count=149
*.plsql_optimize_level=2
*.plsql_code_type=’NATIVE’
PL/SQL native compilation is recommended for Oracle Database 10g or 11g based
Oracle E-Business Suite environments such as Release 12. Interpreted mode is supported and can be used with
Oracle E-Business Suite. However, native compilation is recommended in order to maximize runtime performance
and scalability. Compiling PL/SQL units with native compilation takes longer than using interpreted mode, because
of the need to generate and compile the native shared libraries. Hence need through testing before
implementation to Prod.
3. Administration
Recommendation:
BIAS recommends working with Oracle support to validate the invalid objects i.e. if these objects can be ignored as
invalid. If they cannot be ignored, we recommend compiling them accordingly.
Comment:-
GL_ACCESS_SET_LEDGERS INVALID
PJI_TIME_PA_RPT_STR_MV INVALID
Currently two invalid objects will compile and try to work with Oracle incase invalid still exists.
4. Performance
BIAS Recommendation:
Recommends purging the workflow entries older than two to four weeks through the purge batch program in
order maintain optimal performance and storage levels.
Comment:- Minimum age of data to purge, in days needs to be provided by as per business requirements.
5. Performance
BIAS Recommendation:
To purge the log/out file at regular interval to release space in accordance with business requirements. This would
help in enhancing the performance of concurrent managers.
Comment:-Business needs to provide the number of days for which the log/out files should be kept.
(Recommendation is 60-90 days log/out files to be kept).
Already we implemented the purge procedure in UAT. we will implement the same in pre-prod.
6. Performance
To run once every week with gather statistics estimate percent set to 30%.
Comment:- Estimate_percent
Percentage of rows to estimate. If left empty it will default to 10%. The valid range is 0-99. A higher percentage will
be more accurate, but take longer to run. If the object(s) that you are gathering statistics for do not change often
or the object(s) has data entered that is very similar you may choose a lower number. However, if the data
changes frequently a larger number entered for this parameter would be recommended to provide a more
accurate representation of your data. Hence this value needs to be determined according to data changes and the
time of run of request.
10% will be sufficient at present and as load increases, we can increase them to 30% particularly if we have the
pl/sql code issues.
7. Performance
BIAS Recommendation:
Gc cr block busy and gc buffer busy aquire indicates significant write/write contention. BIAS recommends
increasing each redo log file size from 1GB to 2 GB. Control files/redo log files are also advised to be moved into a
different ASM disk for better performance.
Comment:-Redo Log File size needs to be determined by log file switch and peak time ,data volume and
transactions. Large Redo Size has an impact on the recovery incase required hence we need to be monitor the log
file switches and data volumes for determining the right size of redo log file. Control files and redo log files are kept
in same mount point as per DC standards.
8. Performance
BIAS Recommendation:
Rebuilding the below mentioned indexes is recommended.
APPLSYS.AD_FILES_U2
CN.CN_OBJECTS_N7
CN.CN_OBJECTS_N8
APPLSYS.AD_SNAPSHOT_FILES_N1
APPLSYS.JDR_ATTRIBUTES_U1
APPLSYS.AD_PATCH_COMMON_ACTIONS_U2
APPLSYS.AD_PATCH_RUN_BUG_ACTIONS_U2
APPLSYS.AD_SNAPSHOT_FILES_U2
CN.CN_OBJECTS_N5
APPLSYS.JDR_ATTRIBUTES_N1
CommentL-We can rebuild the indexes but most of the indexes mentioned above is of AD patch and snapshots.
We do not see any fragmented blocks for these indexes and as this is a fresh installation at this point , there is no
need for the index rebuild.
9. Architecture / Performance
BIAS Recommendation
Distribute the Concurrent Managers among the 2 nodes and assign primary /secondary nodes for all the
concurrent managers.
Set s_conc status to disabled in xml file on lasfsxs016q and run autoconfig.
Comment:- This is a optional parameter. if we set the s_conc status to disabled we cannot start the concurrent
manager.
10. DB Tier
Performance
It is a recommended practice to defragment the below mentioned tables for the better performance of Concurrent
requests.
Defragment the tables periodically to reclaim unused space / improve performance
FND_CONCURRENT_REQUESTS,
FND_CONCURRENT_PROCESSES,
FND_CRM_HISTORY,
FND_ENV_CONTEXT,
FND_TEMP_FILES
FND_LOG_MESSAGES
Comment:-As the instance is currently build this tables are of not of huge size to have fragmentation. This could be
implemented once the instance is being used and tables have grown and need defragmentation
Comment:-All patches as per Section 2.1.2 of Note 437878.1 are applied to the instance. This is a required post
step as per the installation document.
Required patches are applied and some of the above patches belongs to 11i not for R12.
Comment:-This patch is available for 11i version, We are on R12.1.3 hence this patch is not applicable.
Oracle Recommendation:
-1 JVM can normally handle up to 100-250 forms users
Recommended settings
-Xmx256M -Xms128M -XX:MaxPermSize=128M -XX:NewRatio=2
Comment:-We are waiting for number of users for Prod and Pre-Prod to set this parameters.
As there is less free memory, we have not implemented this yet. we will check it out and implement the same.
15. Memory
BIAS Recommendation:
Increase the memory to 50 GB on each apps servers since the expected concurrent users is around 1000.
Comment:-
[applmgr@lasfsxs015q ~]$
Since free memory is (2418 MB and 12108 MB(Buffers/cache) hence there is sufficient memory).
16. Architecture
BIAS Recommendation Mid-Term
Shut down the applications on all
nodes and execute the below
EXEC
FND_CONC_CLONE.SETUP_CLEAN;
COMMIT;
EXIT;
Cooment:-As per architecture the PCP is enabled on all three node s but currently enabled for two nodes hence it
should be reviewed further before disabling
17. Architecture
BIAS Recommendation:
Application binaries on a nfs mount point is not recommended and they should moved to SAN.
rw,intr,bg,hard,timeo=600,wsize=32768,rsize=32768,nfsvers=3,tcp,nolock,acregmin=0,acregmax=0
DB Oracle Home
rw,bg,intr,hard,timeo=600,acregmin=0,acregmax=0,wsize=32768,rsize=32768,nfsvers=3,tcp
Data Files
rw,bg,nointr,hard,proto=tcp,vers=3,rsize=32768,wsize=32768 ( For RAW Devices)
s_forms_time: This is for the time for the stale process.So, at present it is better to have the default value of 30.
s_sesstimeout: As we have a less memory , we can have the default value at present.
20. Performance
Below mentioned performance Mid-Term
related bugs/patches are
recommended to be applied to 11g
DB Home.
1.RDBMS 7533387 DYNAMIC
SAMPLING IS NOT HAPPENING FOR
PL/SQL TABLES. 11.1.0.7.0