Friday Dec 28, 2012

Solaris Tips : CPU Cache Sizes, Changing System Date

Tip #1: Finding the CPU cache sizes from Solaris operating environment

Use the prtpicl utility to list out system configuration, and look for the cache sizes within that output.


$ /usr/sbin/prtpicl -v |grep cache
              :l1-icache-size    0x10000
              :l1-icache-line-size       0x40
              :l1-icache-associativity   0x2
              :l1-dcache-size    0x10000
              :l1-dcache-line-size       0x40
              :l1-dcache-associativity   0x2
              :l2-cache-size     0x500000
              :l2-cache-line-size        0x100
              :l2-cache-associativity    0xa

[Updated 01/14/13] The above output was gathered from an M4000 system that has SPARC64 VII processors.

Recent update releases of Solaris 10 and 11 show the prtpicl reported cache sizes in decimal numbers.

Here is a slightly improved prtpicl command that filters out unwanted output. (Courtesy: Georg)

/usr/sbin/prtpicl -v -c cpu | egrep "^ +cpu|ID|cache"

Tip #2: Changing the System Date

Use date to change the system date. For example, to set the system date to March 9, 2008 08:15 AM, run the following command. Syntax: date mmddHHMMyy

#date 0309081508

Sun Mar 9 08:15:03 PST 2008

Friday Nov 23, 2012

emca fails with "Database instance is unavailable" though available

The following example shows the symptoms of failure, and the exact error message.

$ emca -repos create

Password for SYSMAN user:  

Do you wish to continue? [yes(Y)/no(N)]: Y
Nov 19, 2012 10:33:42 AM oracle.sysman.emcp.DatabaseChecks \
WARNING: ORA-01034: ORACLE not available

Nov 19, 2012 10:33:42 AM oracle.sysman.emcp.DatabaseChecks \
Database instance is unavailable. Fix the ORA error thrown and 
run EM Configuration Assistant again.

Some of the possible reasons may be : 

1) Database may not be up. 
2) Database is started setting environment variable ORACLE_HOME 
with trailing '/'. Reset ORACLE_HOME and bounce the database. 

For eg. Database is started setting environment variable 
ORACLE_HOME=/scratch/db/ . Reset ORACLE_HOME=/scratch/db  and bounce 
the database.


Ensure that the ORACLE_HOME is pointing to the right location in $ORACLE_HOME/bin/emca file.

Rather than installing from scratch, if ORACLE_HOME was copied over from another location, likely it results in wrong location for ORACLE_HOME in several Enterprise Manager (EM) specific scripts and files. It usually happens when the directory structure on the target machine is not identical to the structure on the original/source machine, including the top level directory location where Oracle RDBMS was installed properly using the installer.

Monday Oct 15, 2012

Consolidating Oracle E-Business Suite R12 on Oracle's SPARC SuperCluster

An Optimized Solution for Oracle E-Business Suite (EBS) R12 12.1.3 is now available on

    The Oracle Optimized Solution for Oracle E-Business Suite

This solution was centered around the engineered system, SPARC SuperCluster T4-4. Check the business and technical white papers along with a bunch of relevant useful resources online at the above optimized solution page for EBS.

What is an Optimized Solution?

Oracle's Optimized Solutions are designed, tested and fully documented architectures that are tuned for optimal performance and availability. Optimized solutions are NOT pre-packaged, fully tuned, ready-to-install software bundles that can be downloaded and installed. An optimized solution is usually a well documented architecture that was thoroughly tested on a target platform. The technical white paper details the deployed application architecture along with various observations from installing the application on target platform to its behavior and performance in highly available and scalable configurations.

Oracle E-Business Suite R12 Use Case

Multiple E-Business Suite R12 12.1.3 application modules were tested in this optimized solution -- Financials (online - oracle forms & web requests), Order Management (online - oracle forms & web req uests) and HRMS (online - web requests & payroll batch). The solution will be updated with additional application modules, when they are available.

Oracle Solaris Cluster is responsible for the high availability portion of the solution.

Performance Data

For the sake of completeness, test results were also documented in the optimized solution white paper. Those test results are mainly for educational purposes only. They give good sense of application behavior under the circumstances the application was tested. Since the major focus of the optimized solution is around highly available and scalable configurations, the application was configured to me et those criteria. Hence the documented test results are not directly comparable to any other E-Business Suite performance test results published by any vendor including Oracle. Such an attempt may lead to skewed, incorrect conclusions.

Questions & Requests

Feel free to direct your questions to the author of the white papers. If you are a potential customer who would like to test a specific E-Business Suite application module on any non-engineered syste m such as SPARC T4-X or engineered system such as SPARC SuperCluster, contact Oracle Solution Center.

Monday Sep 24, 2012

E-Business Suite : Role of CHUNK_SIZE in Oracle Payroll

Different batch processes in Oracle Payroll flow have the ability to spawn multiple child processes (or threads) to complete the work in hand. The number of child processes to fork is controlled by the THREADS parameter in APPS.PAY_ACTION_PARAMETERS view.

THREADS parameter

The default value for THREADS parameter is 1, which is fine for a single-processor system but not optimal for the modern multi-core multi-processor systems. Setting the THREADS parameter to a value equal to or less than the total number of [virtual] processors available on the system may improve the performance of payroll processing. However on the down side, since multiple child processes operate against the same set of payroll tables in HR schema, database may experience undesired consequences such as buffer busy waits and index contention, which results in giving up some of the gains achieved by using multiple child processes/threads to process the work. Couple of other action parameters, CHUNK_SIZE and CHUNK_SHUFFLE, help alleviate the database contention.


Set a value for THREADS parameter as shown below.




(I am not aware of any maximum value for THREADS parameter)

CHUNK_SIZE parameter

The size of each commit unit for the batch process is controlled by the CHUNK_SIZE action parameter. In other words, chunking is the act of splitting the assignment actions into commit groups of desired size represented by the CHUNK_SIZE parameter. The default value is 20, and each thread processes one chunk at a time -- which means each child process inserts or processes 20 assignment actions at any time.

When multiple threads are configured, each thread picks up a chunk to process, completes the assignment actions and then picks up another chunk. This is repeated until all the chunks are exhausted.

It is possible to use different chunk sizes in different batch processes. During the initial phase of processing, CHUNK_SIZE number of assignment actions are inserted into relevant table(s). When multiple child processes are inserting data at the same time into the same set of tables, as explained earlier, database may experience contention. The default value of 20 is mostly optimal in such a case. Experiment with different values for the initial phase by +/-10 for CHUNK_SIZE parameter and observe the performance impact. A larger value may make sense during the main processing phase. Again experimentation is the key in finding the suitable value for your environment. Start with a large value such as 2000 for the chunk size, then increment or decrement the size by 500 at a time until an optimal value is found.


Set a value for CHUNK_SIZE parameter as shown below.




CHUNK_SIZE action parameter accepts a value that is as low as 1 or as high as 16000.


By default, chunks of assignment actions are processed sequentially by all threads - which may not be a good thing especially given that all child processes/threads performing similar actions against the same set of tables almost at the same time. By saying not a good thing, I mean to say that the default behavior leads to contention in the database (in data blocks, for example).

It is possible to relieve some of that database contention by randomizing the processing order of chunks of assignment actions. This behavior is controlled by the CHUNK SHUFFLE action parameter. Chunk processing is not randomized unless explicitly configured.


Set chunk shuffling as shown below.




Finally I recommend checking the following document out for additional details and additional pay action tunable parameters that may speed up the processing of Oracle Payroll.
    My Oracle Support Doc ID: 226987.1 Oracle 11i & R12 Human Resources (HRMS) & Benefits (BEN) Tuning & System Health Checks

Also experiment with different combinations of parameters and values until the right set of action parameters and values are found for your deployment.

Friday Aug 03, 2012

Enabling 2 GB Large Pages on Solaris 10

Few facts:

  • - 8 KB is the default page size on Solaris 10 and 11 as of this writing
  • - both hardware and software must have support for 2 GB large pages
  • - SPARC T4 hardware is capable of supporting 2 GB pages
  • - Solaris 11 kernel has in-built support for 2 GB pages
  • - Solaris 10 has no default support for 2 GB pages
  • - Memory intensive 64-bit applications may benefit the most from using 2 GB pages


OS: Solaris 10 8/11 (Update 10) or later
Hardware: SPARC T4. eg., SPARC T4-1, T4-2 or T4-4

Steps to enable 2 GB large pages on Solaris 10:

  1. Install the latest kernel patch or ensure that 147440-04 or later was installed

  2. Add the following line to /etc/system and reboot
    • set max_uheap_lpsize=0x80000000

  3. Finally check the output of the following command when the system is back online
    • pagesize -a

    % pagesize -a
    8192		<-- 8K
    65536		<-- 64K
    4194304		<-- 4M
    268435456	<-- 256M
    2147483648	<-- 2G
    % uname -a
    SunOS jar-jar 5.10 Generic_147440-21 sun4v sparc sun4v

Also See:

Sunday Jul 29, 2012

[OID] ldap_modify: Failed to find member in mandatory or optional attribute list

A sample LDAP entry and the resulting error message are shown below. The objective is simple - adding a new member (employee) to an existing group (Administrators).

% cat assigngrp.ldif

dn: cn=Administrators,ou=groups,ou=entapp
changetype: modify
add: member
member: cn=emp1234,ou=people,ou=entapp

% ldapmodify -p 3060 -h localhost -D "cn=orcladmin" -w passwd -f assigngrp.ldif
add member:
modifying entry cn=Administrators,ou=groups,ou=entapp
ldap_modify: Object class violation
ldap_modify: additional info: Failed to find member in mandatory or \
     optional attribute list.

The above error message is a generic one. It would have been nice had it shown the expected and actual inputs as part of the error. However it gave us a hint that the object class was violated. In this example, the group "Administrators" was created under object class groupOfUniqueNames.

% ldapsearch -p 3060 -h localhost -b "ou=groups,ou=entapp" -A "(objectclass=*)"

RFC 4519 for Lightweight Directory Access Protocol (LDAP) requires the uniqueMember attribute within the groupOfUniqueNames object class. An excerpt from the original RFC:

3.6.  'groupOfUniqueNames'

      ( NAME 'groupOfUniqueNames'
         SUP top
         MUST ( uniqueMember $
               cn )
         MAY ( businessCategory $
               seeAlso $
               owner $
               ou $
               o $
               description ) )

Going back to the issue in hand, the "add" attribute must be uniqueMember, not member, in "modify" LDAP entry. That's the object class violation in this case. Now the fix to the issue is obvious.

The modified entry and the output from Oracle Internet Directory's ldapmodify command are shown below.

% cat assigngrp.ldif

dn: cn=Administrators,ou=groups,ou=entapp
changetype: modify
add: uniqueMember
uniqueMember: cn=emp1234,ou=people,ou=entapp

$ ldapmodify -p 3060 -h localhost -D "cn=orcladmin" -w passwd -f assigngrp.ldif
add uniqueMember:
modifying entry cn=Administrators,ou=groups,ou=entapp
modify complete

Though the above example was derived from an Oracle Internet Directory (OID) environment, the problem and the solution are applicable to all environments running LDAP servers.

Tuesday Jun 12, 2012

Oracle E-Business Suite Tip : SQL Tracing


Attempts to enable SQL tracing from concurrent request form fails with error:

Function not available to this responsibility.
Change Responsibilities or contact your System Administrator


Switch responsibility to "System Administrator". Navigate to System -> Profiles, and query for "%Diagnostics% ("Utilities : Diagnostics")". Once found the profile, change its value to "Yes". Restart web browser and try enabling SQL trace again.

Tuesday May 08, 2012

OBIEE 11g: Resolving Presentation Services Startup Failure


Starting Presentation Services fail with the error:

[OBIPS] [ERROR:1] [] [] [ecid: ] [tid: ] Authentication Failure.
Odbc driver returned an error (SQLDriverConnectW).
State: 08004.  Code: 10018.  [NQODBC] [SQL_STATE: 08004] [nQSError: 10018] Access for the requested connection is refused.
[nQSError: 43113] Message returned from OBIS.
[nQSError: 43126] Authentication failed: invalid user/password. (08004)[[

Also connecting to the metadata repository (RPD) in online mode fails with similar error.

Looking through the BI server log, nqserver.log, you may find an error message similar to the following:

[OracleBIServerComponent] [ERROR:1] [] [] [ecid: 0001J1LfUetFCC3LVml3ic0000pp000000] [tid: 1] 
[13026] Error in getting roles from BI Security Service:	 
'Error Message From BI Security Service: [nQSError: 46164] HTTP Server returned 404 (Not Found) for URL .' ^M


  • Connect to WebLogic Server (WLS) Console -> Deployments. Ensure that all deployed components are in 'Active' state.

  • If any of the components is in 'Prepared' state, select that application and then click on "start servicing all requests"

  • Restart BI Server and Presentation Services

In some cases, the following additional step might be needed to resolve the issue.

  • Access the Enterprise Manager Fusion Middleware control: http://<host.domain>:port/em

  • Navigate to Business Intelligence -> coreapplication

  • 'Capacity Management' tab -> 'Scalability' sub-tab

  • Click on 'Lock and Edit Configuration' button

  • Enter the IP address in the 'Listen Address' field

  • Click on 'Activate Changes' followed by 'Release Configuration' buttons

  • Restart BI Server and Presentation Services

Also check these My Oracle Support (MOS) documents for more clues and information.

1387283.1 Authentication failed: invalid user/password
1251364.1 Error: "[nQSError: 10018] Access .. Refused. [nQSError: 43126] Authentication Failed .." when Installing OBIEE 11g
1410233.1.1 How To Bind Components / Ports To A Specific IP Address On Multiple Network Interface (NIC) Machines

Friday Apr 27, 2012

Solaris Volume Manager (SVM) on Solaris 11

SVM is not installed on Solaris 11 by default.

# metadb
-bash: metadb: command not found

# /usr/sbin/metadb
-bash: /usr/sbin/metadb: No such file or directory

Install it using pkg utility.

# pkg info svm
pkg: info: no packages matching the following patterns you specified are
installed on the system.  Try specifying -r to query remotely:


# pkg info -r svm
          Name: storage/svm
       Summary: Solaris Volume Manager
   Description: Solaris Volume Manager commands
      Category: System/Core
         State: Not installed
     Publisher: solaris
       Version: 0.5.11
 Build Release: 5.11
Packaging Date: October 19, 2011 06:42:14 AM 
          Size: 3.48 MB
          FMRI: pkg://solaris/storage/svm@0.5.11,5.11-

# pkg install storage/svm
           Packages to install:   1
       Create boot environment:  No
Create backup boot environment: Yes
            Services to change:   1

DOWNLOAD                                  PKGS       FILES    XFER (MB)
Completed                                  1/1     104/104      1.6/1.6

PHASE                                        ACTIONS
Install Phase                                168/168 

PHASE                                          ITEMS
Package State Update Phase                       1/1 
Image State Update Phase                         2/2 

# which metadb

This time metadb may fail with a different error.

# metadb
metadb: <HOST>: /dev/md/admin: No such file or directory

Check if md.conf exists.

# ls -l  /kernel/drv/md.conf 
-rw-r--r--   1 root     sys          295 Apr 26 15:07 /kernel/drv/md.conf

Dynamically re-scan md.conf so the device tree gets updated.

# update_drv -f md

# ls -l  /dev/md/admin
lrwxrwxrwx   1 root root 31 Apr 20 10:12 /dev/md/admin -> ../../devices/pseudo/md@0:admin

# metadb
metadb: <HOST>: there are no existing databases

Now Solaris Volume Manager is ready to use.

#  metadb -f -a c0t5000CCA00A5A7878d0s0

# metadb
        flags           first blk       block count
     a        u         16              8192          /dev/dsk/c0t5000CCA00A5A7878d0s0

Friday Mar 30, 2012

Resolving "PLS-00201: identifier 'DBMS_SYSTEM.XXXX' must be declared" Error

Here is a failure sample.

SQL> set serveroutput on
SQL> alter package APPS.FND_TRACE compile body;

Warning: Package Body altered with compilation errors.

SQL> show errors

-------- -----------------------------------------------------------------
235/6    PL/SQL: Statement ignored
235/6    PLS-00201: identifier 'DBMS_SYSTEM.SET_EV' must be declared

By default, DBMS_SYSTEM package is accessible only from SYS schema. Also there is no public synonym created for this package. So, the solution is to create the public synonym and grant "execute" privilege on DBMS_SYSTEM package to all database users or a specific user.


SQL> CREATE PUBLIC SYNONYM dbms_system FOR dbms_system;

Synonym created.


Grant succeeded.

- OR -


Grant succeeded.

SQL>  alter package APPS.FND_TRACE compile body;

Package body altered.

Note that merely granting execute privilege is not enough -- creating the public synonym is as important to resolve this issue.


Benchmark announcements, HOW-TOs, Tips and Troubleshooting


« July 2015