Saturday, July 30, 2022

Apply R12.2 April 2022 CPU patch

We decided to apply April 2022 CPU to EBS. To make patching consistent in both db side and EBS side, DBA applied 2022 April database PSU together with database 19c upgrade. The document for applying April 2022 CPU is 2856621.1. I followed below steps and order to apply it. As always, first of all identify which patches are required and download all patch files to a central location, such as /s46/path/to/CPU_April22

1. Confirm only patch 30399994 was applied:
SQL> select * from ad_bugs where bug_number in (31856779, 30399994);

30399994 Patch R12.ATG_PF.C.delta.9 (Application Technology Family)
31856779 Patch R12.ATG_PF.C.delta.10

2. Identify and decide 9 EBS patches will be applied in this time:
CPU patch:
33782739 RELEASE 12.2: CPU PATCH FOR APR 2022

Post Patching:
33908208:R12.FWK.C (12.2.10 CONSOLIDATED PATCH FOR APR CPU 2022 FIXES FOR FWK)
33207251:R12.FND.C (Application Object Library. Applied as part of 33782739)
32980025:R12.AK.C (Application Framework)
33286000:R12.OAM.C (Oracle Applications Manager)
33625264:R12.OWF.C (Workflow)

Oracle support recommended when troubleshooting oacore (optional):
34005885:R12.FWK.C ORACLE APPL FRAMEWORK (FWK) RELEASE 12.2.10 BUNDLE 9
33659063:R12.FWK.C Image Patch for 12.2.10 Post Consolidated Patch Product OAF
33973100:R12.FND.C - CONNECTION LEAK _PAGES._JSP._FND.__CLOSE._JSPSERVICE 

Patch for a bug when re-creating AZ tables (Doc ID 2816670.1):
p31791566_R12.AZ.C_R12_GENERIC.zip

$ cd $PATCH_TOP

$ cp /s46/path/to/CPU_April22/EBS/*.zip .      (9 zip files)

unzip p33782739_12.2.0_R12_LINUX.zip

unzip p33908208_R12.FWK.C_R12_GENERIC.zip
unzip p32980025_R12.AK.C_R12_GENERIC.zip
unzip p33286000_R12.OAM.C_R12_GENERIC.zip
unzip p33625264_R12.OWF.C_R12_GENERIC.zip

unzip p33659063_R12.FWK.C_R12_GENERIC.zip
unzip p33973100_R12.FND.C_R12_GENERIC.zip
unzip p34005885_R12.FWK.C_R12_GENERIC.zip

unzip p31791566_R12.AZ.C_R12_GENERIC.zip

$ echo $FILE_EDITION
run 

Note: adgrants.sql from patch 33782739 is an older version than the one in $APPL_TOP/admin

$ adop phase=apply apply_mode=downtime patches=33782739 patchtop=$NE_BASE/EBSapps/patch
                                                                                      <== it may take one hour
Post-patch steps (from README):
a) Run autoConfig in RUN file system on all Apps-tie nodes
$ ./adautocfg.sh     
b) $ perl $AD_TOP/bin/admkappsutil.pl
c) Run autoConfig on database node       (Note: I skipped this step)
$ adconfig.sh contextfile=$ORACLE_HOME/.../appsutil/$CONTEXT_NAME.xml

$ adop phase=apply apply_mode=downtime patches=33908208,34005885,33659063,33973100,32980025,33286000,33625264  patchtop=$NE_BASE/EBSapps/patch

$ adop phase=apply apply_mode=downtime patches=31791566 patchtop=$NE_BASE/EBSapps/patch

-- Verify all 10 patches were applied to all nodes
SQL> SELECT adb.bug_number, aas.name appl_top_name, adb.creation_date, adb.language,
decode(ad_patch.is_patch_applied('R12',aas.appl_top_id,adb.bug_number,adb.language),'EXPLICIT','APPLIED','NOT_APPLIED','NOT APPLIED') status
FROM ad_bugs adb,
(select distinct appltop_id appl_top_id, node_name name from ad_adop_sessions 
           where node_name in (select node_name from ADOP_VALID_NODES) ) aas
where adb.bug_number in (
'33782739',
'33908208',
'34005885',
'33659063',
'33973100',
'33207251',
'32980025',
'33286000',
'33625264',
'31791566'
) order by creation_date desc, adb.bug_number,aas.name,adb.language;

3. Run latest ETCC (patch 17537119). This will help in identifying and confirming what patches will be applied.

4. Apply patches for Oracle WebLogic Server (WLS) 10.3.6.0

$ cd $FMW_HOME/utils/bsu/cache_dir                       
$ cp /s46/path/to/CPU_April22/WLS/*.zip .                   (7 .zip files)

$ mv p33800106_1036_Generic.zip ../.
$ cd ..
$ ls
$ unzip p33800106_1036_Generic.zip      (from 33845432)
$ chmod +x bsu_update.sh
$ bsu.sh -version
$ ./bsu_update.sh install           <== it runs very quick.
Installing Smart Update V5...
Updating bsu modules
Update was successful.
$ bsu.sh -version

$ cd cache_dir

$ unzip -o p33791826_1036_Generic.zip  (33946345 WLS PATCH SET UPDATE 10.3.6.0.220419)
$ unzip -o p13845626_10360220419_Generic.zip  (33812339 CONSOLIDATED FMW FIXES)
$ unzip -o p13964737_10360220419_Generic.zip  (zip file from 33812339)
$ unzip -o p33560682_1036_Generic.zip                (from 33812339)
$ unzip -o p31241365_1036_Linux-x86-64.zip      (from 33812339)
$ unzip -o p31042881_1036_Generic.zip                (from 33812339)

$ cd ..
$ ./bsu.sh -prod_dir=$FMW_HOME/wlserver_10.3 -status=applied -verbose -view | egrep -i "T64V|UCI6|N3YF|9UNH|CW7X|7BIA"

-- it detects conflicts    
$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=T64V -prod_dir=$FMW_HOME/wlserver_10.3
Checking for conflicts..
Conflict(s) detected - resolve conflict condition and execute patch installation again
Conflict condition details follow:
Patch T64V is mutually exclusive and cannot coexist with patch(es): YVDZ,1YWL,HJT5

./bsu.sh -remove -patchlist=YVDZ -prod_dir=$FMW_HOME/wlserver_10.3

./bsu.sh -remove -patchlist=HJT5 -prod_dir=$FMW_HOME/wlserver_10.3
   
./bsu.sh -remove -patchlist=1YWL -prod_dir=$FMW_HOME/wlserver_10.3

$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=T64V -prod_dir=$FMW_HOME/wlserver_10.3
Checking for conflicts..
No conflict(s) detected
Installing Patch ID: T64V..
Result: Success

$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=UCI6 -prod_dir=$FMW_HOME/wlserver_10.3
Checking for conflicts..
No conflict(s) detected
Installing Patch ID: UCI6.
Result: Success

$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=N3YF -prod_dir=$FMW_HOME/wlserver_10.3
Checking for conflicts..
No conflict(s) detected
Installing Patch ID: N3YF..
Result: Success

-- it detects a conflict
$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=9UNH -prod_dir=$FMW_HOME/wlserver_10.3
Checking for conflicts..
Conflict(s) detected - resolve conflict condition and execute patch installation again
Conflict condition details follow:
Patch 9UNH is mutually exclusive and cannot coexist with patch(es): DN1F

$ ./bsu.sh -remove -patchlist=DN1F -prod_dir=$FMW_HOME/wlserver_10.3
Checking for conflicts..
No conflict(s) detected
Removing Patch ID: DN1F..
Result: Success

$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=9UNH -prod_dir=$FMW_HOME/wlserver_10.3

$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=CW7X -prod_dir=$FMW_HOME/wlserver_10.3 

$ ./bsu.sh -install -patch_download_dir=$FMW_HOME/utils/bsu/cache_dir -patchlist=7BIA -prod_dir=$FMW_HOME/wlserver_10.3 

$ ./bsu.sh -prod_dir=$FMW_HOME/wlserver_10.3 -status=applied -verbose -view | egrep -i "T64V|UCI6|N3YF|9UNH|CW7X|7BIA"
Patch ID:               T64V
PatchContainer:    T64V.jar
Patch ID:               UCI6
PatchContainer:    UCI6.jar
                   33791826 , Patch ID :T64V) in the environment
Patch ID:               N3YF
PatchContainer:    N3YF.jar
                   U (Patch Number:33791826 , Patch ID :T64V ) in the environme
Patch ID:               9UNH
PatchContainer:    9UNH.jar
Patch ID:               CW7X
PatchContainer:    CW7X.jar
Patch ID:               7BIA
PatchContainer:    7BIA.jar

5. apply 4 patches for Oracle Fusion Middleware (FMW) 11.1.1.9 OHS - Web Tier

$ echo $IAS_ORACLE_HOME  
$ export ORACLE_HOME=$IAS_ORACLE_HOME
$ echo $ORACLE_HOME
$RUN_BASE/FMW_Home/webtier
$ export PATH=$IAS_ORACLE_HOME/OPatch:$PATH
$ which opatch
$RUN_BASE/FMW_Home/webtier/OPatch/opatch

cd /s46/path/to/CPU_April22/OH          (7 zip files)

Note: when all .zip files are saved in a shared location, only unzip them once for all nodes/instances
$ unzip p33963904_111190_Linux-x86-64.zip  
$ cd 33963904

$ opatch apply -jre $ORACLE_HOME/jdk/jre
Oracle Interim Patch Installer version 11.1.0.12.9
Copyright (c) 2022, Oracle Corporation.  All rights reserved.
... ...
Patching component oracle.ohs2, 11.1.1.9.0...
RollbackSession removing interim patch '31047338' from inventory
OPatch back to application of the patch '33963904' after auto-rollback.

Patching component oracle.ohs2, 11.1.1.9.0...
Verifying the update...
Patch 33963904 successfully applied

$ cd ..
$ unzip p32287205_111190_Linux-x86-64.zip
$ cd 32287205
$ opatch apply
... ...
Patching component oracle.ldap.rsf, 11.1.1.9.0...
Patching component oracle.rdbms.rsf, 11.1.0.7.0...
RollbackSession removing interim patch '31304503' from inventory
OPatch back to application of the patch '32287205' after auto-rollback.
Patching component oracle.ldap.rsf, 11.1.1.9.0...
Patching component oracle.rdbms.rsf, 11.1.0.7.0...
Verifying the update...
Patch 32287205 successfully applied

$ cd ..
$ unzip p32928416_111190_Linux-x86-64.zip
$ cd 32928416
$ opatch apply
... ...
Patching component oracle.opmn, 11.1.1.9.0...
Verifying the update...
Patch 32928416 successfully applied

$ cd ..
$ unzip p33144848_111190_Linux-x86-64.zip
$ cd 33144848
$ opatch apply -jre $ORACLE_HOME/jdk/jre
... ...
Patching component oracle.wlsplugins, 11.1.1.9.0...
Verifying the update...
Patch 33144848 successfully applied

$ opatch lsinventory
Interim patches (17) :

6. apply 2 patches for Oracle Fusion Middleware (FMW) - Oracle Common

$ export ORACLE_HOME=$FMW_HOME/oracle_common
$ export PATH=$ORACLE_HOME/OPatch:$PATH

$ echo $ORACLE_HOME
$FMW_HOME/oracle_common
$ which opatch
$FMW_HOME/oracle_common/OPatch/opatch

$ cd ..
$ unzip p31985571_111190_Generic.zip
$ cd 31985571/oui
$ opatch apply
... ...
Patching component oracle.jrf.adfrt, 11.1.1.9.0...
RollbackSession removing interim patch '30368663' from inventory
OPatch back to application of the patch '31985571' after auto-rollback.
Patching component oracle.jrf.adfrt, 11.1.1.9.0...
Verifying the update...
Patch 31985571 successfully applied

$ cd ../..
$ unzip p26933408_111190_Generic.zip
$ cd 26933408
$ opatch apply
... ...
Patching component oracle.jrf.thirdparty.jee, 11.1.1.9.0...
Verifying the update...
Patch 26933408 successfully applied

$ opatch lsinventory
Interim patches (14) : ...

7. apply 1 patch for Oracle Fusion Middleware 10.1.2.3 - Forms and Reports

Start a new OS session
$ echo $FILE_EDITION
run
$ echo $ORACLE_HOME
$RUN_BASE/EBSapps/10.1.2
$ which opatch
$RUN_BASE/EBSapps/10.1.2/OPatch/opatch

$ cd /aetnas46/oracmprod/R122patch_IFSPGRND/CPU_April22/OH
$ unzip p32922089_101232_LINUX.zip
$ cd 32922089
$ opatch apply
... ...
Removing patch 26825525...
Applying patch 32922089...
OPatch succeeded.

$ opatch lsinventory
44 patches.

8. Run ETCC script again to confirm all required technology patches were applied.

9. Testing and and after signed-off, run FS_CLONE.
$ cd $ADMIN_SCRIPTS_HOME
$ ./adautocfg.sh
$ ./adstrtal.sh
Do a backup on file systems and then run fs_clone. 
$ adop phase=fs_clone

Additional notes:
An Oracle document shows Log4j vulnerabilities are mitigated by April 2022 CPU patches. Seems Oracle collected all remediations into this patch set. Here is the document (as of July 30 2022):

CVE-2021-4104/CVE-2022-23302/CVE-2022-23305/CVE-2022-23307 Advisory for Oracle E-Business Suite (Apache Log4j 1.x Vulnerabilities) (Doc ID 2858304.1)

SYMPTOMS
These vulnerabilities have various conditions of exploitability. For example, to be maliciously exploited vulnerability CVE-2022-23307 requires that the attacker be able to create a malicious log entry and either the attacker or a customer run the Chainsaw component on the targeted system. Because of these various conditions of exploitability and the way the Log4j components are used in Oracle products, the reported CVSS Base Scores for these vulnerabilities do not necessarily reflect the criticality of the issues in the context of their use in an Oracle product. These non-default configurations of Log4j are not used by Oracle E-Business Suite.

CVE-2021-4104 (CVSS 7.5) deserialization of untrusted data in JMSAppender
Vulnerable if: log4j configuration uses JMSAppender (not used by default)
CVE-2022-23302 (CVSS 8.8) deserialization of untrusted data in JMSSink
Vulnerable if: log4j configuration uses JMSSink (not used by default)
CVE-2022-23305 (CVSS 9.8) SQL injection in JDBCAppender
Vulnerable if: log4j configuration uses JDBCAppender (not used by default)
CVE-2022-23307 (CVSS 9.8) Chainsaw GUI logviewer (for log4j's XML output)
Vulnerable if: using chainsaw

PATCHES
The following patches are available to mitigate the Log4j related vulnerabilities. If you find Log4j in other components not listed here, file a service request against the respective components.
Core Oracle E-Business Suite Functionality
Apply the following patches:

For Oracle E-Business Suite Release 12.2.x with WLS 10.3.6 and FMW 11.1.1.9:
o Apply EBS April 2022 CPU Patch 33782739.
o Apply WLS 10.3.6 overlay Patch 33796519, which is included in Patch 33570839, on top of the 10.3.6 January 2022 PSU. This patch is included in Patch 33946345 EBS REHOSTED: 33791826 WLS PATCH SET UPDATE 10.3.6.0.220419.
o Apply WLS 10.3.6 Smart Update V5 Patch 33845432,  which is also included in Patch 33570839 EBS RELEASE 12.2 CONSOLIDATED FMW FIXES FOR JAN 2022 and subsequent Critical Patch Updates (CPUs).
o Apply FMW 11.1.1.9 Patch 33960746 (to oracle_common).

The following occurrences of log4j jar files are modified by the aforementioned patches:
FMW_Home/modules/com.bea.core.apache.log4j_1.2.13.jar
FMW_Home/modules/com.bea.core.apache.log4j_1.1.0.0_1-2-15.jar
FMW_Home/wlserver_10.3/server/lib/consoleapp/APP-INF/lib/log4j-1.2.17-16.jar
FMW_Home/wlserver_10.3/server/lib/consoleapp/APP-INF/lib/log4j-1.2.8.jar
FMW_Home/wlserver_10.3/server/lib/wlstestclient.ear/webapp.war/log4j-1.2.8.jar
FMW_Home/patch_wls1036/backup/backup.jar/wlstestclient.ear/webapp.war/log4j-1.2.8.jar
FMW_Home/oracle_common/sysman/archives/fmwctrl/app/em.ear
FMW_Home/oracle_common/sysman/archives/applications/11_1_1_0_0_emcore.war
FMW_Home/oracle_common/sysman/jlib/log4j-core.jar

Any occurrences of log4j jar files in these directories can be manually removed:
All Log4j libraries under the following directories:
o /clone (such as …/EBSapps/comn/clone/)
o /adopclone (such as …/EBSapps/comn/adopclone_xxxxx/)
o /inventory (such as …/EBSapps/10.1.2/ccr/inventory/)
o /patch stage (such as .../bsu/cache_dir/ and .../.patch_storage/)
o FMW_Home/user_projects/domains/EBS_domain/servers/AdminServer/tmp/_WL_user/emcore/28c293/WEB-INF/lib/log4j-core.jar
Shut down the WLS server before removing this file. After a while, the library might come back again which is the patched library.
o Run the following commands to stub out the log4j.jar under the 10.1.2 Oracle home:
     $ cd <ORACLE_HOME>
     $ perl -e print "" > ./sysman/admin/emdrep/lib/log4j.jar
o Run the following commands to stub out the EAR file under the 10.1.2 Oracle home for Oracle E-Business Suite Release 12.2 and 10.1.3 Oracle home for Oracle E-Business Suite Release 12.1:
     $ cd <ORACLE_HOME>
     $ perl -e print "" > ./oc4j/j2ee/oc4j_applications/applications/isqlplus.ear

Our analysis of the following occurrences of log4j v1.x indicates that these can be considered false positives. These instances are false positives since they do not contain the vulnerable classes.

The following Log4j jar files are not impacted by the CVEs listed in this note since they do not contain the vulnerable classes:
10.1.2/sysman/jlib/log4j-core.jar
10.1.2/OPatch/ocm/ocm.zip/core.jar/payload.zip/log4j-core.jar
10.1.2/diagnostics/lib/ojdl-log4j.jar
FMW_Home/Oracle_EBS-app1/ccr/lib/log4j-core.jar
FMW_Home/Oracle_EBS-app1/oui/jlib/lib/log4j-core.jar
FMW_Home/Oracle_OAMWebGate1/oui/jlib/jlib/log4j-core.jar
FMW_Home/webtier/OPatch/ocm/lib/log4j-core.jar
FMW_Home/webtier/OPatch/ocm/ocm.zip/log4j-core.jar
FMW_Home/webtier/OPatch/ocm/ocm.zip/core.jar/payload.zip/log4j-core.jar
FMW_Home/webtier/ccr/lib/log4j-core.jar
FMW_Home/webtier/oui/jlib/jlib/log4j-core.jar
FMW_Home/oracle_common/OPatch/ocm/lib/log4j-core.jar
FMW_Home/oracle_common/OPatch/ocm/ocm.zip/log4j-core.jar
FMW_Home/oracle_common/OPatch/ocm/ocm.zip/core.jar/payload.zip/log4j-core.jar
FMW_Home/oracle_common/modules/oracle.odl_11.1.1/ojdl-log4j.jar
DBhome/sysman/jlib/ocm/log4j-core.jar
DBhome/oui/jlib/jlib/log4j-core.jar
DBhome/oc4j/j2ee/home/applications/ascontrol.ear/ascontrol.war/log4j-core.jar
DBHome/ccr/lib/log4j-core.jar
DBHome/javavm/lib/dbhadoop12101.jar

Vulnerability NOTES: 

In another Oracle document ID 2827804.1 (CVE-2021-44228/CVE-2021-45046/CVE-2021-44832/CVE-2021-45105 Advisory for Oracle E-Business Suite (Apache log4j Vulnerabilities), the Core Oracle E-Business Suite Functionality section asks to apply EBS patch 33672402:R12.TXK.C (in early 2022) to mitigate the vulnerability in Oracle E-Business Suite 12.2 associated with 
CVE-2021-44228, 
CVE-2021-45046, 
CVE-2021-44832, 
CVE-2021-45105

It patches file $COMMON_TOP/java/lib/log4j_core.jar.

Wednesday, July 27, 2022

Add OACORE servers in R12.2

Oracle Doc ID 1905593.1 (Managing Configuration of Oracle HTTP Server and Web Application Services in Oracle E-Business Suite Release 12.2) gives steps on adding oacore process to the instance. See "4.4.1 Adding a new managed server".

Addition of managed servers needs to be done on the RUN file system when there is no active ADOP cycle. During the next adop prepare, the Configuration Change Detector identifies that the addition has been made and the managed servers are automatically synced up from the run file system to the patch file system. The synchronization also gets done when fs_clone is executed.

Most likely, adding CPU and memory to the server are needed to support additional oacore.

1. Backup $CONTEXT_FILE
2. Backup two conf files (in below path or similar. Many conf files resides in this folder)
$ cd  $IAS_ORACLE_HOME/instances/EBS_web_OHS1/config/OHS/EBS_web
mod_wl_ohs.conf
apps.conf

3. Check ADOP session
$ hostname
master_node.domain.com
$ echo $TWO_TASK
EBSDEV

$ adop -status
Enter the APPS password:
Connected.
==============================================================
ADOP (C.Delta.12)
Session Id: 5
Command: status
Output: $NE_BASE/EBSapps/log/adop/5/20220X08_115941/adzdshowstatus.out
===============================================================
Node Name       Node Type  Phase           Status          Started              Finished             Elapsed
--------------- ---------- --------------- --------------- -------------------- -------------------- ------------
2nd_node            slave    FS_CLONE        COMPLETED       2022/02/17 23:05:12  2022/02/18 00:01:37  0:56:25
3rd_node            slave     FS_CLONE        COMPLETED       2022/02/17 23:05:12  2022/02/17 23:59:02  0:53:50
naster_node      master    FS_CLONE        COMPLETED       2022/02/17 23:05:12  2022/02/17 23:39:21  0:34:09

4. Identify port numbers (and how many OACORE server(s) on each node)
$ grep -i oacore_server_ports $RUN_BASE/inst/apps/$CONTEXT_NAME/appl/admin/$CONTEXT_NAME.xml
<oacore_server_ports oa_var="s_oacore_server_ports">oacore_server1:7206</oacore_server_ports>

$ grep -i oacore_server_ports $PATCH_BASE/inst/apps/$CONTEXT_NAME/appl/admin/EBSDEV*.xml
<oacore_server_ports oa_var="s_oacore_server_ports">oacore_server1:7207</oacore_server_ports>

Port 7206 is used for RUN file system and 7207 is used for PATCH file system. Port 7208 is free and is chosen for the new oacore. Also oacore_server1 is running on master node and oacore_server2 is running on 2nd node. Name oacore_server3 as the new server.

5.  Add a new managed server. This will create a managed server and add a new entry to the context file for starting and stopping the new managed server via the adstrtal and adstpall scripts:

$ echo $FILE_EDITION
run
$ echo $APPL_TOP
/u02/app/EBSPROD/fs1/EBSapps/appl    

$ echo $CONTEXT_FILE
$ perl $AD_TOP/patch/115/bin/adProvisionEBS.pl \
        ebs-create-managedserver -contextfile=$CONTEXT_FILE \
        -managedsrvname=oacore_server3 -servicetype=oacore \
        -managedsrvport=7208 -logfile=$APPLRGF/TXK/addMS_oacoreserver3_2022.log
Enter the APPS Schema password:
Enter the WebLogic AdminServer password:

The logfile for this session is located at $APPLRGF/TXK/addMS_oacoreserver3_2022.log
ManagedServer oacore_server3 created successfully.

Note: use the real path and file name for $CONTEXT_FILE in above line

$  egrep -i "Error|Failed|Warning|ORA-" $APPLRGF/TXK/addMS_oacoreserver3_2022.log
Warning: An insecure protocol was used to connect to the
*** ALL THE FOLLOWING FILES ARE REQUIRED FOR RESOLVING RUNTIME ERRORS
Warning: An insecure protocol was used to connect to the

$ grep -i s_oacore_nodes $CONTEXT_FILE
       <oacore_nodes oa_var="s_oacore_nodes">master_node.aetna.com:7206</oacore_nodes>

$ grep -i oacore_server_ports $CONTEXT_FILE
  <oacore_server_ports oa_var="s_oacore_server_ports">oacore_server1:7206,oacore_server3:7208</oacore_server_ports>

By the way, it does not add anything to PATCH file system.
$ grep -i oacore_server_ports $PATCH_BASE/inst/apps/$CONTEXT_NAME/appl/admin/EBSDEV*.xml
<oacore_server_ports oa_var="s_oacore_server_ports">oacore_server1:7207</oacore_server_ports>

Now, starting oacore_server3 shall work:
sh $ADMIN_SCRIPTS_HOME/admanagedsrvctl.sh status oacore_server3
sh $ADMIN_SCRIPTS_HOME/admanagedsrvctl.sh start oacore_server3
sh $ADMIN_SCRIPTS_HOME/admanagedsrvctl.sh status oacore_server3
You are running admanagedsrvctl.sh version 120.14.12020000.12
Enter the WebLogic Admin password:
oacore_server3 is running.

6. Add details of the newly added managed servers into the OHS configuration files mod_wl_ohs.conf and apps.conf on the current node:

$ echo $CONTEXT_FILE
$ perl $FND_TOP/patch/115/bin/txkSetAppsConf.pl \
        -contextfile=$CONTEXT_FILE \
        -configoption=addMS -oacore=node_name.domain.com:7208

Log file shall say mod_wl_ohs.conf and apps.conf were generated successfully.
$ egrep -i "Error|Failed|Warning|ORA-" $APPLRGF/TXK/txkSetAppsConf_*.log

$ cd $ADMIN_SCRIPTS_HOME
$ adapcctl.sh stop
$ adopmnctl.sh status
$ adapcctl.sh start start
$ adopmnctl.sh status

Recycle all services:
$ adstpall.sh apps/appsPWD
$ adstrtal.sh apps/appsPWD
Then, log onto Weblogic console to see two oacore servers are running.

7. Repeat steps on 2nd web/forms node. Skip this on node dedicated for concurrent jobs. 

Before adding more oacore
$ grep -i oacore_server_ports $CONTEXT_FILE
   <oacore_server_ports oa_var="s_oacore_server_ports">oacore_server2:7206</oacore_server_ports>

After adding one more oacore
$ grep -i s_oacore_nodes $CONTEXT_FILE
      <oacore_nodes oa_var="s_oacore_nodes">master_node.aetna.com:7206</oacore_nodes>
Note: it points to master_node on the 2nd node config. This could be the reason for unbalanced number of connections.

$ grep -i oacore_server_ports $CONTEXT_FILE
  <oacore_server_ports oa_var="s_oacore_server_ports">oacore_server2:7206,oacore_server4:7208</oacore_server_ports>

Again, verify it does not do anything on PATCH file system.
$ grep -i oacore_server_ports $PATCH_BASE/inst/apps/$CONTEXT_NAME/appl/admin/EBSDEV*.xml
<oacore_server_ports oa_var="s_oacore_server_ports">oacore_server2:7207</oacore_server_ports>

8.  Summary on oacore name and port:
Node1   RUN  oacore_server1:7206,   oacore_server3:7208
Node2   RUN  oacore_server2:7206,   oacore_server4:7208
After an adop CUTOVER or FS_CLONE, the change will be copied to PATCH file system. Names will be the same but use different ports:
Node1   PATCH  oacore_server1:7207,   oacore_server3:7209
Node2   PATCH  oacore_server2:7207,   oacore_server4:7209

The original idea for adding more oacore servers to our R12.2.10 instance was to address the unbalance of connection loadings between oacore_server1 and oacore_server2. But adding more oacore servers still does not resolve the issue. oacore servers on master node still get much more connections. 
The benefit is that if one OACORE is down or frozen, we can recycle it while the 2nd one is still running. That could reduce downtime on end-users.

UPDATES in 2023:
After we opened a SR with Oracle Support on OACORE connection distribution uneven, Oracle engineer provided an action plan for us to add entries to apps.conf and mod_wl_ohs.conf files which will allow for load balancing of traffic to all OACOREs.

1. Take a backup of your EBS environment
2. Connect to master_node and add oacore_server2 and oacore_server4 to the configuration:

a. Open a terminal window and connect to master node
b. Switch to the apps user
c. Source the apps environment
d. Run the commands below:

$ echo $FILE_EDITION
run
$ echo $APPL_TOP
/u02/app/EBSPROD/fs2/EBSapps/appl     <== a CUSOVER was done after oacore was added 

$ perl $FND_TOP/patch/115/bin/txkSetAppsConf.pl \
-contextfile=<CONTEXT_FILE> \
-configoption=addMS \
-oacore=2nd_node.domain.com:7207

$ perl $FND_TOP/patch/115/bin/txkSetAppsConf.pl \
-contextfile=<CONTEXT_FILE> \
-configoption=addMS \
-oacore=2nd_node.domain.com:7209

e. Restart OHS by running adapcctl.sh

3. Connect to 2nd_node and add oacore_server3 to the configuration:

a. Open a terminal window and connect to 2nd_node
b. Switch to the apps user
c. Source the apps environment
d. Run the below command:

$ perl $FND_TOP/patch/115/bin/txkSetAppsConf.pl \
-contextfile=<CONTEXT_FILE> \
-configoption=addMS \
-oacore=master_node.domain.com:7209

e. Restart OHS by running adapcctl.sh

4. Direct your load balancer team to stop sending web requests to your CONCURRENT MANAGER ONLY node.
5. Retest for the issue

The txkSetAppsConf.pl command should be run after adding apps tier nodes to your EBS environment. This is mentioned in the excerpt below from Doc ID 1383621.1 :
"If any of the managed servers from the newly added node are desired to be part of the cluster configuration on the current node, run the following command to add details of these managed servers into the OHS configuration files mod_wl_ohs.conf and apps.conf on the current node:..."

Before performing action plan from the SR (on master_node), entries in two .conf files: 
$ cd $IAS_ORACLE_HOME/instances/EBS_web_OHS1/config/OHS/EBS_web
$ grep 7209 apps.conf
BalancerMember http://master_node.domain.com:7209/OA_HTML/media
BalancerMember http://master_node.domain.com:7209/OA_HTML/classes
BalancerMember http://master_node.domain.com:7209

$ grep 7209 mod_wl_ohs.conf
WebLogicCluster master_node.domain.com:7207,master_node.domain.com:7209

$ grep 7207 apps.conf
BalancerMember http://master_node.domain.com:7207/OA_HTML/media
BalancerMember http://master_node.domain.com:7207/OA_HTML/classes
BalancerMember http://master_node.domain.com:7207

$ grep 7207 mod_wl_ohs.conf
WebLogicCluster master_node.domain.com:7207,master_node.domain.com:7209

After performing action plan from the SR (on master_node), all entries are balanced with two nodes.

$ grep 7209 apps.conf
   BalancerMember http://master_node.domain.com:7209/OA_HTML/media
   BalancerMember http://2nd_node.domain.com:7209/OA_HTML/media
   BalancerMember http://master_node.domain.com:7209/OA_HTML/classes
   BalancerMember http://2nd_node.domain.com:7209/OA_HTML/classes
   BalancerMember http://master_node.domain.com:7209
   BalancerMember http://2nd_node.domain.com:7209

$ grep 7209 mod_wl_ohs.conf
WebLogicCluster master_node.domain.com:7207,master_node.domain.com:7209,2nd_node.domain.com:7207,2nd_node.domain.com:7209

$ grep 7207 apps.conf
   BalancerMember http://master_node.domain.com:7207/OA_HTML/media
   BalancerMember http://2nd_node.domain.com:7207/OA_HTML/media
   BalancerMember http://master_node.domain.com:7207/OA_HTML/classes
   BalancerMember http://2nd_node.domain.com:7207/OA_HTML/classes
   BalancerMember http://master_node.domain.com:7207
   BalancerMember http://2nd_node.domain.com:7207

$ grep 7207 mod_wl_ohs.conf
WebLogicCluster master_node.domain.com:7207,master_node.domain.com:7209,2nd_node.domain.com:7207,2nd_node.domain.com:7209

Very similar entries in .conf files in 2nd_node, after action plan was performed on 2nd_node.