1. "XXXX is not a valid responsibility for the current user. Please contact your System Administrator."
- First possibility is that this responsibility has an End date for this user. Check user's account.
- After adding a responsibility to a user, the user may receive above message. And the attached responsibility may not appear in the Home Page until the Apache is bounced.
One solution could be to clear the cache:
Navigate to Functional administrator > Core Service > Caching Framework > Global Configuration > Clear All Cache
Another place to check:
Search for profile 'Applications Start Page'. Delete the value set at user level (if any).
Also make sure Workflow Service Components are up and running (OAM -> Workflow Manager (dropdown) -> Service Components):
. Workflow Deferred Agent Listener
. Workflow Java Deferred Agent Listener
2. "Function not available to this responsibility. Change responsibilities or contact System Administrator"
- If it is on a custom form, make sure its path CUSTOM_TOP is added to $INST_TOP/ora/10.1.2/forms/server/default.env file.
- Use submenu item Help --> Diagnostics --> Examine. Then, click the "Examine Fields and variable values" to troubleshoot. Reference Doc ID 1079951.1.
- If this message shows in R12.1.3 while trying to access Help --> Diagnostics, change the profile "Utilities:Diagnostics" to Yes at user or responsibility level. Please see Doc ID 1200743.1.
3. If R12.1 "adformsctl.sh start" gets error, check logs:
$LOG_HOME/ora/10.1.3/j2ee/forms/forms_default_group_1/application.log
$LOG_HOME/ora/10.1.3/j2ee/oacore/oacore_default_group_1/application.log
When the Form service is starting, it checks whether latest formsapp.ear is deployed. If it finds a newer version, Form service will take longer to start and say message:
==========================================
*** Latest formsapp.ear has NOT been deployed ***
Deploying the latest EAR file...
==========================================
See DocID 1210660.1 (Forms Startup Script adformsctl.sh Shows Deployment Message While Starting Forms Services).
4. How to deploy Forms EAR (Enterprise ARchive) file, such as formsapp.ear (under $ORACLE_HOME/forms/j2ee) ?
DocID 397174.1 (Deploying a New EAR File in Oracle Applications Release 12) gives two key steps:
1) Edit file (with a temp password) - $INST_TOP/ora/10.1.3/j2ee/forms/config/system-jazn-data.xml
2) Run deployment - $FND_TOP/bin/txkrun.pl -script=CfgOC4JApp
(if all work, it shall say "No error encountered." at the end)
DocID 1945012.1 also gives a way to manually deploy formsapp.ear (for Oracle Forms ?? vs Oracle EBS R12). But I never try it:
$ dcmctl redeployApplication -f $ORACLE_HOME/forms/j2ee/formsapp.ear -a formsapp -co OC4J_BI_Forms
Friday, September 25, 2015
Use adctrl to stop a worker of adpatch
1. Skip a script when it has been "Running" status for too long
During patching, a program, such as cstpostimportaad.sql, stays in "Running" status forever (Doc ID 1453612.1). To stop and skip it (and you may re-run the program after adpatch finishes):
$ adctrl
==> 4. Tell manager that a worker failed its job
==> 3. Tell worker to quit
==> 6. Restart a worker on the current machine. (Otherwise, worker keeps staying in "Assigned")
2. A worker stuck in "Failed" status. How to stop the worker and skip this program?
Worker Code Context Filename Status
---------- -------- ---------------------- ------------------- --------------
3 Done AutoPatch R120 EGOSILDU.sql Failed
$ adctrl
==> 8 (Note: 8 is NOT listed on the menu)
Now, the status becomes:
3 Done AutoPatch R120 EGOSILDU.sql Skip & restart
==> 6. Restart a worker on the current machine
In another case, after I manually uploaded a failed workflow file to database (Doc ID 2581597.1), I used adcrtl to continue the adpatch session:
==> 8
Enter the worker number(s)/range(s) or 'all' for all workers,
or press [Return] to go back to the menu : 2
Status changed to 'Skip & restart' for worker 2.
Review the messages above, then press [Return] to continue.
Once above steps in adctrl were done, adpatch automatically continued to next jobs.
3. How to re-start "adpatch" from scratch?
Sometimes, you have to start adpatch again to clean all memories kept in previous run, such as change of a schema password in the middle of adpatch session. Even using "3. Tell worker to quit" on all workers, adpatch will still run on OS level. You have to use "kill -9" to stop it.
During patching, a program, such as cstpostimportaad.sql, stays in "Running" status forever (Doc ID 1453612.1). To stop and skip it (and you may re-run the program after adpatch finishes):
$ adctrl
==> 4. Tell manager that a worker failed its job
==> 3. Tell worker to quit
==> 6. Restart a worker on the current machine. (Otherwise, worker keeps staying in "Assigned")
2. A worker stuck in "Failed" status. How to stop the worker and skip this program?
Worker Code Context Filename Status
---------- -------- ---------------------- ------------------- --------------
3 Done AutoPatch R120 EGOSILDU.sql Failed
$ adctrl
==> 8 (Note: 8 is NOT listed on the menu)
Now, the status becomes:
3 Done AutoPatch R120 EGOSILDU.sql Skip & restart
==> 6. Restart a worker on the current machine
In another case, after I manually uploaded a failed workflow file to database (Doc ID 2581597.1), I used adcrtl to continue the adpatch session:
Worker Code Context Filename Status
---------- -------- ---------------------- ------------------- --------------
2 Run AutoPatch R120 POSSPAPM.wft FAILED==> 8
Enter the worker number(s)/range(s) or 'all' for all workers,
or press [Return] to go back to the menu : 2
Status changed to 'Skip & restart' for worker 2.
Review the messages above, then press [Return] to continue.
Once above steps in adctrl were done, adpatch automatically continued to next jobs.
3. How to re-start "adpatch" from scratch?
Sometimes, you have to start adpatch again to clean all memories kept in previous run, such as change of a schema password in the middle of adpatch session. Even using "3. Tell worker to quit" on all workers, adpatch will still run on OS level. You have to use "kill -9" to stop it.
Sunday, September 13, 2015
Re-new SSL certificate file & How to enable SSL
In my environment, two files ewallet.p12 and cwallet.sso are saved in $INST_TOP/certs/Apache. The path is defined in ssl.conf file. Autoconfig will add cwallet.sso as the keystore variable to $CONTEXT_FILE.
Seems to me that certificate file works for the host name or site name. If one server hosts multiple EBS instances, same certificate file can be used for all Apache instances on the host. Certificate expires after a period of time.
Steps to renew certificate files (from .p12 file):
1. Generate the new cert file (renew the file) from certificate tool
Depending on the tool provided by CA (certificate authority) of the company for generating the certificate file, select to Renew cert for the site or host with key size of 2048 kb.
2. Download the .p12 file in my company's tool
Check Yes to include Root Chain and Private key (in PKCS12 format, depending on the tool). Also enter a password to protect the file.
8. verify file cwallet.sso gets created in the same directory $INST_TOP/certs/Apache.
How to enable SSL?
1. Create new certificate file for site ebssitename.domain.com by the certificate tool and get the two files ready in $INST_TOP/certs/Apache (similar to Steps 1 to 9 above).
2. Update 5 entries in Context file on web tiers: active web port, login, external url, webentryhost, webentryurlprotocol
- active web port (s_active_webport) : 80 --> 443
Usually F5 listens to 443, the HTTPS default, and redirects connections to real ssl port, such as 4472. For real port (without F5), makes sure the port (such as 4472) for s_active_webport and s_webssl_port in $CONTEXT_FILE are the same and matches the port in $INST_TOP/ora/10.1.3/Apache/Apache/conf/ssl.conf.
- login (s_login_page): https://ebssitename.domain.com:s_active_webport/OA_HTML/AppsLogin
- external url (s_external_url): https://ebssitename.domain.com:s_active_webport
- webentryhost (s_webentryhost): ebssitename (no change)
- webentryurlprotocol (s_webentryurlprotocol): http --> https
3. It may be necessary to modify ssl.conf in $INST_TOP/ora/10.1.3/Apache/Apache/conf to enable TLS1.x by editing file $FND_TOP/admin/template/ssl_conf_1013.tmp
4. Run auotconfig on all web nodes (this step will update Port number in httpd.conf, etc).
How to disable SSL?
1. Change above 5 variables in $CONTEXT_FILE. or
2. After shurdown apps tier, run
$ txkrun.pl -script=SetAdvCfg \-appsuser=apps -appspass=<Apps Password> -disable=SSL \-s_webport=<Web port>(Replace your Web Port Number)
3. "Running Configuration Wizards from the Command Line in Oracle E-Business Suite Release 12 [ID 1364368.1]", under "Disabling SSL" section.
Reference:
Doc ID 376700.1 - Enabling SSL or TLS in Oracle E-Business Suite Release 12
Seems to me that certificate file works for the host name or site name. If one server hosts multiple EBS instances, same certificate file can be used for all Apache instances on the host. Certificate expires after a period of time.
Steps to renew certificate files (from .p12 file):
1. Generate the new cert file (renew the file) from certificate tool
Depending on the tool provided by CA (certificate authority) of the company for generating the certificate file, select to Renew cert for the site or host with key size of 2048 kb.
2. Download the .p12 file in my company's tool
Check Yes to include Root Chain and Private key (in PKCS12 format, depending on the tool). Also enter a password to protect the file.
3. Copy the new cert file to $INST_TOP/certs/Apache and name it ewallet.p12 (after backing up the old one).
4. Go to $INST_TOP/ora/10.1.3 directory, and source the .env file
Make sure $ORACLE_HOME point to 10.1.3 HOME ( for all infrastructure changes).
5. $ owm (run Oracle Wallet Manager after starting Exceed for XWindow).
NOTES: if run owm from 10.1.2 Oracle Home, it may give an error, such as
owm: line 155: /d2/R12/ab/apps/R1211XB9/apps/tech_st/10.1.2/jdk/jre//bin/java: No such file or directory
6. In OWM GUI, Wallet => Open, navigate to folder $INST_TOP/certs/Apache, click on the path => Open
7. Enter the password (the one entered when re-new and download certificate file) and then click Wallet => check "Auto Login" to enable it, then => Save.
Note in February 2016: If new certificate standard SHA2 (DigiCert) replaces SHA1 (GeoTrust) on generating new .p12 file, OWM may not open file ewallet.p12 with invalid password error. The fix is to apply patch 21845960 (October 2015 CPU Patch) to 10.1.3 ORACLE HOME (for Apache).
Note in February 2016: If new certificate standard SHA2 (DigiCert) replaces SHA1 (GeoTrust) on generating new .p12 file, OWM may not open file ewallet.p12 with invalid password error. The fix is to apply patch 21845960 (October 2015 CPU Patch) to 10.1.3 ORACLE HOME (for Apache).
8. verify file cwallet.sso gets created in the same directory $INST_TOP/certs/Apache.
9. Copy the two files to other nodes of same SITE (if site uses multiple nodes for network load balance).
10. Open a new session to take regular .env file and then run autoconfig in web nodes.
11. Start apps services. Click on the lock icon in the URL, you shall see the new date in certificate file.
12. Copy the two files to other instance(s) on the same node (if applicable).
How to enable SSL?
1. Create new certificate file for site ebssitename.domain.com by the certificate tool and get the two files ready in $INST_TOP/certs/Apache (similar to Steps 1 to 9 above).
2. Update 5 entries in Context file on web tiers: active web port, login, external url, webentryhost, webentryurlprotocol
- active web port (s_active_webport) : 80 --> 443
Usually F5 listens to 443, the HTTPS default, and redirects connections to real ssl port, such as 4472. For real port (without F5), makes sure the port (such as 4472) for s_active_webport and s_webssl_port in $CONTEXT_FILE are the same and matches the port in $INST_TOP/ora/10.1.3/Apache/Apache/conf/ssl.conf.
- login (s_login_page): https://ebssitename.domain.com:s_active_webport/OA_HTML/AppsLogin
- external url (s_external_url): https://ebssitename.domain.com:s_active_webport
- webentryhost (s_webentryhost): ebssitename (no change)
- webentryurlprotocol (s_webentryurlprotocol): http --> https
3. It may be necessary to modify ssl.conf in $INST_TOP/ora/10.1.3/Apache/Apache/conf to enable TLS1.x by editing file $FND_TOP/admin/template/ssl_conf_1013.tmp
4. Run auotconfig on all web nodes (this step will update Port number in httpd.conf, etc).
How to disable SSL?
1. Change above 5 variables in $CONTEXT_FILE. or
2. After shurdown apps tier, run
$ txkrun.pl -script=SetAdvCfg \-appsuser=apps -appspass=<Apps Password> -disable=SSL \-s_webport=<Web port>(Replace your Web Port Number)
3. "Running Configuration Wizards from the Command Line in Oracle E-Business Suite Release 12 [ID 1364368.1]", under "Disabling SSL" section.
Reference:
Doc ID 376700.1 - Enabling SSL or TLS in Oracle E-Business Suite Release 12
UPDATES in 2020: After TLS1.2 is enabled (see https://erpondb.blogspot.com/2020/03/enable-tls12-in-ebs-r121.html), more steps are needed in re-newing ssl certificate because more configuration files need and use new keys from .p12 file. I write a script to extract keys.
#!/bin/bash
# To get new files server.key and opmn.crt, from new cert file ewallet.p12, used by template files
# in $FND_TOP/admin/template/custom
# Assumpation: TLS1.2 was enabled. And, intermediate.crt & ca.crt are company-wide
# and are unchanged from year to year.
# Steps: 1. go to $INST_TOP/ora/10.1.3 to make 10.1.3 the ORACLE_HOME, and then modify the .p12 file by owm
# 2. place this file in $INST_TOP/certs/Apache, and then run it.
#
p12PWD='N0Pe'
echo -n "p12 cert password > "
read p12PWD
if [ $p12PWD != 'N0Pe' ]; then
echo "$p12PWD"
CURRPWD=$PWD
echo $CURRPWD
echo "working on cert files ..."
# Assume two certs ca.crt and intermediate.crt are universal within the company.
# certs files
export PATH=$IAS_ORACLE_HOME/Apache/open_ssl/bin:$PATH
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$IAS_ORACLE_HOME/lib
export OPENSSL_CONF=$IAS_ORACLE_HOME/Apache/open_ssl/bin/openssl.cnf
cd $IAS_ORACLE_HOME/Apache/open_ssl/bin
chmod 755 openssl
which openssl
cd $INST_TOP/certs/Apache
if [ -f 'jyy.key' ]; then
mv jyy.key jyy.key_OLD
fi
if [ -f 'server.key' ]; then
mv server.key server.key_OLD
fi
if [ -f 'server.crt' ]; then
mv server.crt server.crt_OLD
fi
if [ -f 'opmn.crt' ]; then
mv opmn.crt opmn.crt_OLD
fi
if [ -f 'allcerts.jyy' ]; then
mv allcerts.jyy allcerts.jyy_OLD
fi
openssl pkcs12 -in ewallet.p12 -out jyy.key -nocerts -passout pass:'change1t' -password pass:$p12PWD
openssl rsa -in jyy.key -out server.key -passin pass:'change1t'
# below line assumes the SERVER cert is the 1st position in ewallet.p12.
# If it is not, need to manually copy it from allcerts.jyy by next command line.
openssl pkcs12 -in ewallet.p12 -clcerts -nokeys -password pass:$p12PWD | sed -ne '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p;/-END CERTIFICATE-/q' > server.crt
# Use this to get details of 3 certs. Note allcerts.jyy is just for
# troubleshooting/comparison purpose when needed.
openssl pkcs12 -in ewallet.p12 -out allcerts.jyy -nokeys -password pass:$p12PWD
#
cat server.crt intermediate.crt ca.crt > opmn.crt
#
# Not sure if the order in below output file is always correct
openssl pkcs12 -in ewallet.p12 -clcerts -nokeys -password pass:$p12PWD | sed -ne '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p' > opmn.crt_2nd
# opmn.crt and opmn.crt_2nd shall be identical except the key order in the files
# echo "Use below line to check difference. should have nothing"
# diff opmn.crt opmn.cer_2nd
ls -altr
echo $INST_TOP/certs/Apache
cd $CURRPWD
else
echo "password seems wrong."
exit 1
fi
#!/bin/bash
# To get new files server.key and opmn.crt, from new cert file ewallet.p12, used by template files
# in $FND_TOP/admin/template/custom
# Assumpation: TLS1.2 was enabled. And, intermediate.crt & ca.crt are company-wide
# and are unchanged from year to year.
# Steps: 1. go to $INST_TOP/ora/10.1.3 to make 10.1.3 the ORACLE_HOME, and then modify the .p12 file by owm
# 2. place this file in $INST_TOP/certs/Apache, and then run it.
#
p12PWD='N0Pe'
echo -n "p12 cert password > "
read p12PWD
if [ $p12PWD != 'N0Pe' ]; then
echo "$p12PWD"
CURRPWD=$PWD
echo $CURRPWD
echo "working on cert files ..."
# Assume two certs ca.crt and intermediate.crt are universal within the company.
# certs files
export PATH=$IAS_ORACLE_HOME/Apache/open_ssl/bin:$PATH
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$IAS_ORACLE_HOME/lib
export OPENSSL_CONF=$IAS_ORACLE_HOME/Apache/open_ssl/bin/openssl.cnf
cd $IAS_ORACLE_HOME/Apache/open_ssl/bin
chmod 755 openssl
which openssl
cd $INST_TOP/certs/Apache
if [ -f 'jyy.key' ]; then
mv jyy.key jyy.key_OLD
fi
if [ -f 'server.key' ]; then
mv server.key server.key_OLD
fi
if [ -f 'server.crt' ]; then
mv server.crt server.crt_OLD
fi
if [ -f 'opmn.crt' ]; then
mv opmn.crt opmn.crt_OLD
fi
if [ -f 'allcerts.jyy' ]; then
mv allcerts.jyy allcerts.jyy_OLD
fi
openssl pkcs12 -in ewallet.p12 -out jyy.key -nocerts -passout pass:'change1t' -password pass:$p12PWD
openssl rsa -in jyy.key -out server.key -passin pass:'change1t'
# below line assumes the SERVER cert is the 1st position in ewallet.p12.
# If it is not, need to manually copy it from allcerts.jyy by next command line.
openssl pkcs12 -in ewallet.p12 -clcerts -nokeys -password pass:$p12PWD | sed -ne '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p;/-END CERTIFICATE-/q' > server.crt
# Use this to get details of 3 certs. Note allcerts.jyy is just for
# troubleshooting/comparison purpose when needed.
openssl pkcs12 -in ewallet.p12 -out allcerts.jyy -nokeys -password pass:$p12PWD
#
cat server.crt intermediate.crt ca.crt > opmn.crt
#
# Not sure if the order in below output file is always correct
openssl pkcs12 -in ewallet.p12 -clcerts -nokeys -password pass:$p12PWD | sed -ne '/-BEGIN CERTIFICATE-/,/-END CERTIFICATE-/p' > opmn.crt_2nd
# opmn.crt and opmn.crt_2nd shall be identical except the key order in the files
# echo "Use below line to check difference. should have nothing"
# diff opmn.crt opmn.cer_2nd
ls -altr
echo $INST_TOP/certs/Apache
cd $CURRPWD
else
echo "password seems wrong."
exit 1
fi
UPDATES:
I tried to use orapki to get server.crt file from ewallet.p12 (in R12.1.3). It was very close to make it work.
$ cd $INST_TOP/ora/10.1.3
$ ls
$ . $CONTEXT_NAME.env
$ which orapki <= 10.1.3
$ cd /path/to/Apache
$ mv server.crt server.crt_BK_good
$ orapki wallet display -wallet /path/to/Apache/ewallet.p12
Enter wallet password: xxxxxx
Requested Certificates:
Subject: CN=site_name.domian.com,O=company_name,L=city,ST=AZ,C=US
User Certificates:
Trusted Certificates:
Subject: CN=site_name.domian.com,O=company_name,L=city,ST=AZ,C=US
Subject: CN=company_name Secure CA2,O=company_name,C=US
Subject: CN=DigiCert High Assurance EV Root CA,OU=www.digicert.com,O=DigiCert Inc,C=US
$ orapki wallet export -wallet /path/to/Apache/ewallet.p12 -dn 'CN=site_name.domian.com,O=company_name,L=city,ST=AZ,C=US' -cert server.crt
$ chmod +r server.crt
$ more server.crt
$ diff server.crt server.crt_BK_good
... ... <== the length of lines is different
\ No newline at end of file <== also this message
I used openssl to view their contents:
$ export PATH=$IAS_ORACLE_HOME/Apache/open_ssl/bin:$PATH
$ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$IAS_ORACLE_HOME/lib
$ export OPENSSL_CONF=$IAS_ORACLE_HOME/Apache/open_ssl/bin/openssl.cnf
$ which openssl
~/$TWO_TASK/tech_st/10.1.3/Apache/open_ssl/bin/openssl
$ openssl version
OpenSSL 1.0.2q 20 Nov 2018
$ openssl x509 -in server.crt -noout -text > server.txt
$ openssl x509 -in server.crt_BK_good -noout -text > server.txt_good
$ diff server.txt server.txt_good <== two text files are identical!
$ openssl x509 -noout -modulus -in server.crt | openssl md5
(stdin)= 637615c9fcc4c8945817b4941bbff708
$ openssl rsa -noout -modulus -in server.key | openssl md5
(stdin)= 637615c9fcc4c8945817b4941bbff708
But when using this server.crt to start Apache, it gets error (which is misleading). :
$ ./adapcctl.sh start
You are running adapcctl.sh version 120.7.12010000.2
Starting OPMN managed Oracle HTTP Server (OHS) instance ...
sslSecureInit: SSL_CTX_use_certificate_chain_file($INST_TOP/certs/Apache/server.key): failed
4148582144:error:0906D066:PEM routines:PEM_read_bio:bad end line:pem_lib.c:805:
4148582144:error:140DC009:SSL routines:SSL_CTX_use_certificate_chain_file:PEM lib:ssl_rsa.c:708:
opmnctl: opmn start failed.
opmnctl: opmn is not running.
adapcctl.sh: exiting with status 0
The errors match Doc ID 2721500.1 (OPMN Start Fails with Error : SslSecureInit: SSL_CTX_use_certificate_chain_file Failed). But the solutions do not work. It may have a format or "new line" issue. Or, orapki only works well after version 11.1.1.7.0 (in R12.2)
Monday, September 7, 2015
Upgrade database of EBS R12.1.3 to Oracle 12c (12.1.0.2)
Steps to upgrade EBS database from 11.2.0.2 to 12.1.0.2:
0. Make sure pre-requisites for patches of Step 2 were applied:
select * from ad_bugs where bug_number in (
'8502056', -- R12.AD.B.DELTA.2. for 12923944
'7651091', -- R12.ATG_PF.B.DELTA.2. for 12923944
'6636108', -- AutoConfig patch (Doc ID 387859.1, Section 3.2)
'8919489', -- AutoConfig patch (Doc ID 387859.1, Section 3.1)
'9239089', -- R12.AD.B.DELTA.3 for 16289505
'8919491' -- R12.ATG_PF.B.delta.3 for 16289505
-- '18039691' -- AFTER DB upgrade. for 8796558 on WMS
);
1. Pre-step - Apply patch 11791600:R12.AR.B to address error " Failed on compiling library ARXRWAPP.pll " from applying below 5 patches in Step 2. This could be optional to some environments.
2. Apply 5 required patches by "Interoperability Notes EBS 12.0 or 12.1 with RDBMS 12cR1 (Doc ID 1524398.1)".
p12923944_R12.FND.B_R12_LINUX.zip
p13473483_R12.IBE.B_R12_GENERIC.zip (for iStore. no pre-requisite)
p16289505_12.1.0_R12_LINUX.zip
p19558309_R12.AR.B_R12_LINUX.zip
p8796558_R12.WMS.B_R12_LINUX.zip
NOTES:
(1) Verified that instance has all pre-requisites, including RPC patches for R12.1.3. In my instance, no RPC patch was applied before and so patch 18843706 is not needed.
SQL> select * from ad_bugs where bug_number in (
'17774755', -- RPC1
'19030202', -- RPC2
'20203366' -- RPC3);
"For the 12.1.1 instance, there is no reason to apply that patch since you can not apply a RPC patch unless you are on 12.1.3". See Doc ID 1563799.1 (FAQ: E-Business Suite Financials Recommended Patch Collection - RPC ) for more on RPC patches.
UPDATES in April 2018:
21236633 -- R12.1.3+ RPC4. See Doc ID 2053709.1
22644544 -- R12.1.3+ RPC5. See Doc ID 2152266.1
(2) Also verified ORACLE_HOME patch was applied before.
In my instance, patch 6400501 was applied to 10.1.2 Oracle Home on apps server.
(3) It works fine for me to merge above 6 patches together and then apply it at one time.
(4) After 6 patches are applied and before doing database 12c upgrade in next Step 3, start Apps services to do a quick check-out. All shall work.
3. DBAs upgrade database from 11.2.0.2 to 12.1.0.2 (see patch list in Doc ID 761570.1 appendix A) and also apply July2015 PSU patch (Doc ID 21150768.8).
Patch Conflict patch
17694377 12.1.0.2 Patchset (files p17694377_121020_platform_1of8.zip,
p17694377_121020_platform_2of8.zip,
p17694377_121020_platform_7of8.zip)
21150768 Combo of 12.1.0.2.4 OJVM PSU and 12.1.0.2.4 DB PSU (Jul 2015)
6880880 Latest OPatch
X 19382851 Already included in 12.1.0.2.4 PSU
X 19393542 Already included in 12.1.0.2.4 PSU
X 19627012 Already included in 12.1.0.2.4 PSU
X 19649152 Already included in 12.1.0.2.4 PSU
20204035 12.1.0.2.0 version. No conflicts
After all, it includes the patch numbers for the 12.1.0.2 release itself, the one-offs required for R12.1.3 EBS and ones that are included in the July PSU
$ $ORACLE_HOME/OPatch/opatch lsinventory | grep applied
Patch 21068507 : applied on Wed Jul 29 11:22:52 EDT 2015 (part of patch 21150768)
Patch 20204035 : applied on Wed Jul 29 10:40:05 EDT 2015
Patch 20831110 : applied on Tue Jul 28 16:04:35 EDT 2015 (part of patch 21150768)
After dbupgdiag.sql script as part of pre 12c upgrade steps, it may report invalid objects in SYS and SYSTEM schemas and utlrp.sql did not compile them.
Invalid views V$LOCKS_WITH_COLLISIONS ,V$LOCK_ACTIVITY & V$LOCK_ELEMENT are only for Oracle 6 compatibility and can be dropped if invalid after migration. Follow the below document for dropping them: Doc ID 837667.1 (Invalid SYS Objects Found On 9.2 Before Upgrade To 10.2.0.1). For invalid package system.OSS, follow Doc ID 1087534.1 (Package Body OSS and OSSADMIN In SYSTEM Schema Are Invalid After Database Upgrade).
Also, dropped MGDSYS schema in 11.2.0.3 as pre upgrade step. After post-upgrade step adconfig.sh on database server, created MGDSYS which helped compile package WMS_EPC_PVT body from invalid to valid. (Before MGDSYS was re-created, compiling WMS_EPC_PVT got same error as described in Doc ID 2198041.1. But after re-creation of MGDSYS, the compiling worked without EBS patch 19666945 even WMS is Inactive status, not installed, in the instance).
Invalid packages QPR_MAINTAIN_AW (Doc ID 1136003.1) and ZPB_BUILD_METADATA (Doc ID 1951646.1) can be ignored, and sys.V$KXSBD can be dropped in 12c if it exists.
NOTES: Patch Set Updates (PSU) are cumulative patches that include both the security fixes and priority fixes. The key with PSUs is they are minor version upgrades (e.g., 12.1.0.2.0 to 12.1.0.2.4). Once a PSU is applied, only PSUs can be applied in future quarters until the database is upgraded to a new base version.
4. Follow Doc ID 1524398.1 to do the next steps
Get scripts/files ready for DBA tasks:
(step 21) cp -p $APPL_TOP/admin/adgrants.sql $APPLPTMP/.
(step 22) cp -p $AD_TOP/patch/115/sql/adctxprv.sql $APPLPTMP/.
(step 25) cp -p $APPL_TOP/admin/adstats.sql $APPLPTMP/.
(step 27) Generate new file appsutil.zip for DB server (Doc ID 387859.1, Step 3.1 B)
$ cd $AD_TOP/bin
$ perl admkappsutil.pl
$ cp $INST_TOP/admin/out/appsutil.zip $APPLPTMP/.
Step 24: Validate Workflow ruleset
$ cd $FND_TOP/patch/115/sql
$ sqlplus apps/appPWD @wfaqupfix.sql APPLSYS APPS
Step 26. Deregister the current database server.
Note: it is not necessary, if " select * from fnd_nodes; " output is fine!
SQL> exec fnd_conc_clone.setup_clean;
Step 27. Run adconfig.sh on database server
Step 29: Apply post-upgrade WMS patch 18039691 -- Seems optional (if WMS module is not used)
Step 32: Re-create grants and synonyms
$ adadmin
--> 4 --> 2.
Step 35 Log on to Oracle E-Business Suite with the "System Administrator" responsibility.
Click Requests > Run > Single Request and the OK button. Enter the following parameters:
Request Name = Workflow Directory Services User/Role Validation
Batch Size = 10000
Fix dangling users = Yes
Add missing user/role assignments = Yes
Update WHO columns in WF tables = No
Click "OK" and "Submit".
Notes: it may take more than one hour!
NOTE: After all above patches and steps, it got error with db login issue on running concurrent reports cross database link (defined by ILM archiving & purging tool). I had to apply apps patch16946854 to resolve the errors:
REP-0167: Database connect strings are not permitted in USERID parameter. Use TNSnames alias to specify the database.
REP-0501: Unable to connect to the specified database.
ORA-01017: invalid username/password; logon denied
Extra step: e-Business Suite CPU patches:
Apply e-business suite July 2015 CPU patch 20953340 (Doc ID 2013117.1):
1. apply pre-requisite patch 19273341 R12.BNE.B.delta.4
2. run SQL> @adgrants.sql apps
as SYSDBA (newer version is /20953340/admin/adgrants.sql)
3. apply CPU patch 20953340
Note: for this patch, "adpatch" will run autoconfig during the patching process. All shall be smooth.
I merged 19273341, 20953340 and 16946854 (for db login fix), and then applied it successfully on EBS instances, after the database 12c upgrade and database Jul 2015 PSU patching.
UPDATES in February 2016:
1) During the busiest time in the month, 12.1.0.2 database got hang and no login was allowed. There were many processes waiting on “LGWR worker group ordering”. It did not give ORA- errors described in Doc ID 1957710.1 (but DEADLOCK might be seen if a system state dump was performed). We applied the workaround "_use_single_log_writer"=TRUE to disable a performance feature in 12c for using multiple log writers.
2) After database upgrade to 12.1.0.2, " AR XLAACCUP module: Accounting Program " concurrent job failed very quick with error:
An internal error occurred. Please inform your system administrator or support representative that:
An internal error has occurred in the program xla_ae_lines_pkg.AccountingReversal. ORA-01555: snapshot too old: rollback segment number 28
The fix is EBS patch 19591608. We tried workaround in 2009636.1 to set temp_undo_enabled = FALSE. But it did not fix the problem. Then set it back to TRUE before applying patch 19591608.
3) If 12.1.0.2 database for EBS R12.1 gets below UTL_HTTP error, apply 10gAS Oct2015 CUP patch 21845960 to 10.1.3.5 ORACLE_HOME (for Apache) as the fix (see more detail).
SQL> SELECT utl_http.request('https://sitename.domain.com:443', NULL ,'file:/path/to/wallet', NULL) from dual;
SELECT utl_http.request('https://sitename.domain.com:443', NULL ,'file:/path/to/wallet', NULL) from dual
*
ERROR at line 1:
ORA-29273: HTTP request failed
ORA-29048: Unexpected message received.
ORA-06512: at "SYS.UTL_HTTP", line 1491
ORA-06512: at line 1
See here for setting up Wallet file on database host
4) We also applied R12.1 patch 21950151 to address the performance issue on " ARRGTA: Journal Entries Report ". Note this patch seems having impact on Process Lockboxes report (ARLPLB).
5) After database was upgraded to 12.1.0.2, and database PSU and apps CPU were applied, EBS R12.1.3 got an Import Rules issue. Patch 21149920, which changes only 3 XLA packages that are related to Import functionality, was applied. It fixed the issue.
0. Make sure pre-requisites for patches of Step 2 were applied:
select * from ad_bugs where bug_number in (
'8502056', -- R12.AD.B.DELTA.2. for 12923944
'7651091', -- R12.ATG_PF.B.DELTA.2. for 12923944
'6636108', -- AutoConfig patch (Doc ID 387859.1, Section 3.2)
'8919489', -- AutoConfig patch (Doc ID 387859.1, Section 3.1)
'9239089', -- R12.AD.B.DELTA.3 for 16289505
'8919491' -- R12.ATG_PF.B.delta.3 for 16289505
-- '18039691' -- AFTER DB upgrade. for 8796558 on WMS
);
1. Pre-step - Apply patch 11791600:R12.AR.B to address error " Failed on compiling library ARXRWAPP.pll " from applying below 5 patches in Step 2. This could be optional to some environments.
2. Apply 5 required patches by "Interoperability Notes EBS 12.0 or 12.1 with RDBMS 12cR1 (Doc ID 1524398.1)".
p12923944_R12.FND.B_R12_LINUX.zip
p13473483_R12.IBE.B_R12_GENERIC.zip (for iStore. no pre-requisite)
p16289505_12.1.0_R12_LINUX.zip
p19558309_R12.AR.B_R12_LINUX.zip
p8796558_R12.WMS.B_R12_LINUX.zip
NOTES:
(1) Verified that instance has all pre-requisites, including RPC patches for R12.1.3. In my instance, no RPC patch was applied before and so patch 18843706 is not needed.
SQL> select * from ad_bugs where bug_number in (
'17774755', -- RPC1
'19030202', -- RPC2
'20203366' -- RPC3);
"For the 12.1.1 instance, there is no reason to apply that patch since you can not apply a RPC patch unless you are on 12.1.3". See Doc ID 1563799.1 (FAQ: E-Business Suite Financials Recommended Patch Collection - RPC ) for more on RPC patches.
UPDATES in April 2018:
21236633 -- R12.1.3+ RPC4. See Doc ID 2053709.1
22644544 -- R12.1.3+ RPC5. See Doc ID 2152266.1
(2) Also verified ORACLE_HOME patch was applied before.
In my instance, patch 6400501 was applied to 10.1.2 Oracle Home on apps server.
(3) It works fine for me to merge above 6 patches together and then apply it at one time.
(4) After 6 patches are applied and before doing database 12c upgrade in next Step 3, start Apps services to do a quick check-out. All shall work.
3. DBAs upgrade database from 11.2.0.2 to 12.1.0.2 (see patch list in Doc ID 761570.1 appendix A) and also apply July2015 PSU patch (Doc ID 21150768.8).
Patch Conflict patch
17694377 12.1.0.2 Patchset (files p17694377_121020_platform_1of8.zip,
p17694377_121020_platform_2of8.zip,
p17694377_121020_platform_7of8.zip)
21150768 Combo of 12.1.0.2.4 OJVM PSU and 12.1.0.2.4 DB PSU (Jul 2015)
6880880 Latest OPatch
X 19382851 Already included in 12.1.0.2.4 PSU
X 19393542 Already included in 12.1.0.2.4 PSU
X 19627012 Already included in 12.1.0.2.4 PSU
X 19649152 Already included in 12.1.0.2.4 PSU
20204035 12.1.0.2.0 version. No conflicts
After all, it includes the patch numbers for the 12.1.0.2 release itself, the one-offs required for R12.1.3 EBS and ones that are included in the July PSU
$ $ORACLE_HOME/OPatch/opatch lsinventory | grep applied
Patch 21068507 : applied on Wed Jul 29 11:22:52 EDT 2015 (part of patch 21150768)
Patch 20204035 : applied on Wed Jul 29 10:40:05 EDT 2015
Patch 20831110 : applied on Tue Jul 28 16:04:35 EDT 2015 (part of patch 21150768)
After dbupgdiag.sql script as part of pre 12c upgrade steps, it may report invalid objects in SYS and SYSTEM schemas and utlrp.sql did not compile them.
Invalid views V$LOCKS_WITH_COLLISIONS ,V$LOCK_ACTIVITY & V$LOCK_ELEMENT are only for Oracle 6 compatibility and can be dropped if invalid after migration. Follow the below document for dropping them: Doc ID 837667.1 (Invalid SYS Objects Found On 9.2 Before Upgrade To 10.2.0.1). For invalid package system.OSS, follow Doc ID 1087534.1 (Package Body OSS and OSSADMIN In SYSTEM Schema Are Invalid After Database Upgrade).
Also, dropped MGDSYS schema in 11.2.0.3 as pre upgrade step. After post-upgrade step adconfig.sh on database server, created MGDSYS which helped compile package WMS_EPC_PVT body from invalid to valid. (Before MGDSYS was re-created, compiling WMS_EPC_PVT got same error as described in Doc ID 2198041.1. But after re-creation of MGDSYS, the compiling worked without EBS patch 19666945 even WMS is Inactive status, not installed, in the instance).
Invalid packages QPR_MAINTAIN_AW (Doc ID 1136003.1) and ZPB_BUILD_METADATA (Doc ID 1951646.1) can be ignored, and sys.V$KXSBD can be dropped in 12c if it exists.
NOTES: Patch Set Updates (PSU) are cumulative patches that include both the security fixes and priority fixes. The key with PSUs is they are minor version upgrades (e.g., 12.1.0.2.0 to 12.1.0.2.4). Once a PSU is applied, only PSUs can be applied in future quarters until the database is upgraded to a new base version.
4. Follow Doc ID 1524398.1 to do the next steps
Get scripts/files ready for DBA tasks:
(step 21) cp -p $APPL_TOP/admin/adgrants.sql $APPLPTMP/.
(step 22) cp -p $AD_TOP/patch/115/sql/adctxprv.sql $APPLPTMP/.
(step 25) cp -p $APPL_TOP/admin/adstats.sql $APPLPTMP/.
(step 27) Generate new file appsutil.zip for DB server (Doc ID 387859.1, Step 3.1 B)
$ cd $AD_TOP/bin
$ perl admkappsutil.pl
$ cp $INST_TOP/admin/out/appsutil.zip $APPLPTMP/.
Step 24: Validate Workflow ruleset
$ cd $FND_TOP/patch/115/sql
$ sqlplus apps/appPWD @wfaqupfix.sql APPLSYS APPS
Step 26. Deregister the current database server.
Note: it is not necessary, if " select * from fnd_nodes; " output is fine!
SQL> exec fnd_conc_clone.setup_clean;
Step 27. Run adconfig.sh on database server
Step 29: Apply post-upgrade WMS patch 18039691 -- Seems optional (if WMS module is not used)
Step 32: Re-create grants and synonyms
$ adadmin
--> 4 --> 2.
Step 35 Log on to Oracle E-Business Suite with the "System Administrator" responsibility.
Click Requests > Run > Single Request and the OK button. Enter the following parameters:
Request Name = Workflow Directory Services User/Role Validation
Batch Size = 10000
Fix dangling users = Yes
Add missing user/role assignments = Yes
Update WHO columns in WF tables = No
Click "OK" and "Submit".
Notes: it may take more than one hour!
NOTE: After all above patches and steps, it got error with db login issue on running concurrent reports cross database link (defined by ILM archiving & purging tool). I had to apply apps patch16946854 to resolve the errors:
REP-0167: Database connect strings are not permitted in USERID parameter. Use TNSnames alias to specify the database.
REP-0501: Unable to connect to the specified database.
ORA-01017: invalid username/password; logon denied
Extra step: e-Business Suite CPU patches:
Apply e-business suite July 2015 CPU patch 20953340 (Doc ID 2013117.1):
1. apply pre-requisite patch 19273341 R12.BNE.B.delta.4
2. run SQL> @adgrants.sql apps
as SYSDBA (newer version is /20953340/admin/adgrants.sql)
3. apply CPU patch 20953340
Note: for this patch, "adpatch" will run autoconfig during the patching process. All shall be smooth.
I merged 19273341, 20953340 and 16946854 (for db login fix), and then applied it successfully on EBS instances, after the database 12c upgrade and database Jul 2015 PSU patching.
UPDATES in February 2016:
1) During the busiest time in the month, 12.1.0.2 database got hang and no login was allowed. There were many processes waiting on “LGWR worker group ordering”. It did not give ORA- errors described in Doc ID 1957710.1 (but DEADLOCK might be seen if a system state dump was performed). We applied the workaround "_use_single_log_writer"=TRUE to disable a performance feature in 12c for using multiple log writers.
2) After database upgrade to 12.1.0.2, " AR XLAACCUP module: Accounting Program " concurrent job failed very quick with error:
An internal error occurred. Please inform your system administrator or support representative that:
An internal error has occurred in the program xla_ae_lines_pkg.AccountingReversal. ORA-01555: snapshot too old: rollback segment number 28
The fix is EBS patch 19591608. We tried workaround in 2009636.1 to set temp_undo_enabled = FALSE. But it did not fix the problem. Then set it back to TRUE before applying patch 19591608.
3) If 12.1.0.2 database for EBS R12.1 gets below UTL_HTTP error, apply 10gAS Oct2015 CUP patch 21845960 to 10.1.3.5 ORACLE_HOME (for Apache) as the fix (see more detail).
SQL> SELECT utl_http.request('https://sitename.domain.com:443', NULL ,'file:/path/to/wallet', NULL) from dual;
SELECT utl_http.request('https://sitename.domain.com:443', NULL ,'file:/path/to/wallet', NULL) from dual
*
ERROR at line 1:
ORA-29273: HTTP request failed
ORA-29048: Unexpected message received.
ORA-06512: at "SYS.UTL_HTTP", line 1491
ORA-06512: at line 1
See here for setting up Wallet file on database host
4) We also applied R12.1 patch 21950151 to address the performance issue on " ARRGTA: Journal Entries Report ". Note this patch seems having impact on Process Lockboxes report (ARLPLB).
5) After database was upgraded to 12.1.0.2, and database PSU and apps CPU were applied, EBS R12.1.3 got an Import Rules issue. Patch 21149920, which changes only 3 XLA packages that are related to Import functionality, was applied. It fixed the issue.
Subscribe to:
Posts (Atom)