Preclone

[latifa]

Hi Erman,

This post is the continuation of the previous post on migration to RAC.

I tried to apply the preclone on dbTier and AppsTier.
preclone of dbTier worked fine no errors. However AppsTier failed with the following error:

START: Creating WLS config archive.
Script Executed in 1501 milliseconds, returning status 255
ERROR: Script failed, exit code 255

I found the following link : http://srinath-appsdba.blogspot.fr/2015/08/oracle-ebs-r122x-pre-clone-error-and-fix.html

We should restart correctly our services and retry again. OK I tried to do that. But even starting failed with the following errors: adnodemgrctl.sh: exiting with status 1

adnodemgrctl.txt reveals nodemanager.log : weblogic.nodemanager.common.ConfigException: The domain 'EBS_domain_DBOFINP1' at 'null' was not registered in the nodemanager.domains file and dynamic domain registration is not supported. Please register the domain in the nodemanager.domains file.
        at weblogic.nodemanager.server.NMServer.getDomainManager(NMServer.java:259)
        at weblogic.nodemanager.server.Handler.handleDomain(Handler.java:235)
        at weblogic.nodemanager.server.Handler.handleCommand(Handler.java:117)
        at weblogic.nodemanager.server.Handler.run(Handler.java:73)
        at java.lang.Thread.run(Thread.java:724)

My question, can we register this domain ? put a line in the file nodemanager.domains to solve this issue ?

This server is the master one (udkfinp03) and we made its function as concurrent managers server. so disabled all the otehr services. How can the preclone asks for weblogic ?

Restarting the other nodes of weblogic and forms also failed:
ERROR: Skipping startup of oacore_server2 since the AdminServer is down. AdminServer needs to be up before starting/stopping the managed servers.
ERROR: Skipping startup of oaea_server1 since the AdminServer is down. AdminServer needs to be up before starting/stopping the managed servers.
ERROR: Skipping startup of oafm_server2 since the AdminServer is down. AdminServer needs to be up before starting/stopping the managed servers.
ERROR: Skipping startup of forms_server2 since the AdminServer is down. AdminServer needs to be up before starting/stopping the managed servers.
ERROR: Skipping startup of forms-c4ws_server2 since the AdminServer is down. AdminServer needs to be up before starting/stopping the managed servers.


Any Advise please.

Thanks in advance Erman

Latifa

[ErmanArslansOracleBlog]

What is written in "<MiddlewareHome>\wlserver_10.3\common\nodemanager\nodemanager.domains"?

Is the base domain for EBS pointed there?
If not, write it and retry.
This is only thing that I can suggest you right now for this.
Ref: weblogic.nodemanager.common.ConfigException: nodemanager.domains file and dynamic domain registration is not supported (Doc ID 1530774.1)

"For AdminServer needs to be up before starting/stopping the managed servers" ->
is it really so? I mean are you sure that this node is only for conc managers?
Check fnd_nodes and context file of this node..

[ErmanArslansOracleBlog]

Note that, you have another alternative for getting rid of this.
You are doing an in-place operation, an in place postclone execution is your purpose right?
Then, give a try and run adcfgclone.pl directly without preclone.

[latifa]

Hi Erman,

I try to do cloning in place.

perl adcfgclone.pl appsTier.

However, I have the following error message:

Below Oracle Homes are already registered in the inventory:

/opt/application/V12/fs2/FMW_Home/Oracle_EBS-app1
/opt/application/V12/fs2/FMW_Home/webtier
/opt/application/V12/fs2/FMW_Home/oracle_common

ensure that the above Oracle Homes are unregistered from the inventory "/opt/application/V12/oraInventory/ContentsXML/inventory.xml" and re-run adcfgclone.pl script.

I tried in $ORACLE_HOME/oui/bin:
./runInstaller.sh -detachhome ORACLE_HOME=/opt/application/V12/fs2/FMW_Home/oracle_common
and
./runInstaller -detachhome ORACLE_HOME=/opt/application/V12/fs2/FMW_Home/oracle_common

errors.

Which instruction I must use? Also what is the difference between detachhome and unregister ?

Thanks in advance
Latifa

[latifa]

Hi Erman,

I used this :

http://ebs2ofa.blogspot.fr/2014/08/cloning-cheat-list-ebs-11i-r121-r122.html

it's running ...

Let see

[latifa]

Hi Erman,

I left only EBSapps directory (rename the others) even orainventory created a empty one.

Still failing:

Enter value for 1: Enter value for 2: Enter value for 3: Enter value for 4: ERROR:
ORA-12505: TNS:listener does not currently know of SID given in connect
descriptor

Any advice.

Regards,
Latifa

[ErmanArslansOracleBlog]

Did you check the log file?
adcfgclone may not create the tnsnames.ora files.
execution of adgentns.pl may not be succesful..
check the adcfgclone's log file and see where it is failing.. which script or scripts..
The error you are getting may be a result, not the cause.

[ErmanArslansOracleBlog]

Also send me the following info:

EBS version (including the last digit)?
RAC or single instance?

[latifa]

Hi erman,

please find attached my log file of cloning.

SQL> select release_name from apps.fnd_product_groups;

RELEASE_NAME
--------------------------------------------------
12.2.4

SQL> select host_name from gv$instance ;

HOST_NAME
----------------------------------------------------------------
prdc1pdb01
prdc1pdb02

RCloneApplyAppstier_03281712.log

Thank you in advance for your help.

Regards

[ErmanArslansOracleBlog]

Concantrating on this one ->

getConnectionUsingAppsJDBCConnector() -->
    APPS_JDBC_URL=''
    Trying to get connection using SID based connect descriptor
getConnection() -->
    sDbHost    : prdc1pdb01-scan
    sDbDomain  : dalkia.net
    sDbPort    : 1521
    sDbSid     : DBOFINP1
    sDbUser    : APPS
    Trying to connect using SID...
getConnectionUsingSID() -->
    JDBC URL: jdbc:oracle:thin:@prdc1pdb01-scan.dalkia.net:1521:DBOFINP1
WARNING:     Exception occurred: java.sql.SQLException: Listener refused the connection with the following error:
ORA-12505, TNS:listener does not currently know of SID given in connect descriptor
 
    Trying to connect using SID as ServiceName
getConnectionUsingServiceName() -->
    JDBC URL: jdbc:oracle:thin:@(DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=prdc1pdb01-scan.dalkia.net)(PORT=1521))(CONNECT_DATA=(SERVICE_NAME=DBOFINP1)))
    Connection obtained

[ErmanArslansOracleBlog]

Check this note:

12.2 RAC running adcfgclone on Apps node fails with "ORA-12505, TNS:listener does not currently know of SID given in connect descriptor" (Doc ID 2226504.1)

[latifa]

Hi Erman,

I copied clone directory from another server of CP and rerun adcfgclone, it's gone far and it's actually configuring patch file actually. However, there is some errors:

Configuring: Run file system....

AutoConfig has completed with errors for run
You can review AutoConfig portion log later
--> the error was:
Config Cleanup Complete
Errors were encoutered in cleanup, please run adcleansrccfg.pl manually to complete cleanup task:
perl <AD_TOP>/patch/115/bin/adcleansrccfg.pl context=<RUN_CONTEXT_FILE> logfile=<LOG_FILE>

and then continue to clone: logfile

Configuring: Patch file system....
ERROR: Patch file system configuration has failed

Starting to copy the 10.1.2 home from fs2 to fs1 ...
 ... File exists
WARNING: Failed to copy the 10.1.2 home successfully hence exiting
--> normal I didn't suppress all directories before cloning.

My question: what is the role of adcleansrccfg.pl ? Which quind of clean it executes ? Can we run that script before cloning any server ?

Regards,
Latifa


WARNING: Failed to copy the 10.1.2 home successfully hence exiting

[ErmanArslansOracleBlog]

I checked an EBS 12.2 instance and that script is not there.

It may be related with AD and TXK levels.

Send me that script, so I can review.

It seems, it is doing a cleanup.(delete dirty things)

Did you do your cloning activities properly? As you know, there are certain directories  that should be copied and be there, there when you are cloning..

[latifa]

Hi Erman,

adcleansrccfg.pl

cloning failed and even runing : perl $AD_TOP/adcleansrccfg.pl context=run_context_file failed with following errors:

Database connection could not be established. Either the database is down or the APPS credentials supplied are wrong.
*** ALL THE FOLLOWING FILES ARE REQUIRED FOR RESOLVING RUNTIME ERRORS
*** Log File =....

Removing SSO/OID references if present...
stty: : No such device or address
stty: : No such device or address
stty: : No such device or address
stty: : No such device or address

Errors encountered running /opt/application/V12/fs2/EBSapps/appl/fnd/12.0.0/patch/115/bin/txkSetSSOReg.pl

*******FATAL ERROR*******
PROGRAM : (/opt/application/V12/fs2/EBSapps/appl/fnd/12.0.0/bin/txkrun.pl)
TIME    : Wed Mar 29 23:11:14 2017
FUNCTION: TXK::Common::doError [ Level 1 ]
ERRORMSG: ERROR processing <arg> appspass : End of STDIN

WARNING: could not remove OID references
Config Cleanup Complete
Errors were encoutered in cleanup, please check the log for detailsDatabase connection could not be established. Either the database is down or the APPS credentials supplied are wrong.

txkSetOAMReg_Wed_Mar_29_23_11_05_2017.log

I think I should rerun again adcfgclone with suppressing all : orainventory, inst and FMW_Home. Even in patch file system I will suppress EBSapps. I think it will be created again while copied from run file system.

Regards,
Latifa

[ErmanArslansOracleBlog]

Yes.. That script is used in OAM/OID enabled environments.

and YES, you run adcfgclone in a clean state and update me.

[latifa]

Hi Erman,

after failure of clean script, I decided again to clone but only run file system.
cloning was 100% completed but it still running:

WARNING: AutoConfig has completed with  errors .
Please review the AutoConfig section in the logfile. If required, you can re-run AutoConfig from command line  after fixing the problem (I'll see that later when it will finish and rerun autoconf if possible solve issue).
 ......
Failed to truncate ad_nodes_config_status table

WARNING: Failed to update patch_file_system_base to 'CLONED' for ad_adop_session_patches
Skipping update operation for ad_adop_session_patches table
Executing: delete from ADOP_VALID_NODES where node_name not in ('udkfinp03');

still running with delete_adopvalidnodes.sql !

--> what should I do ? cancel it ...

Regards,
latifa

[latifa]

Finally finished ... with following:

ERROR: Failed to delete source nodes from ADOP_VALID_NODES table
Please try to delete the source node details from ADOP_VALID_NODES manually

Skipping the starting of services

INFO :  Rapid Clone completed successfully , but the AutoConfig run recorded  some errors.
Please review the AutoConfig section in the logfile.  and Re-Run the entire cloning cycle , after fixing the problem.

I will check the logfile for autoconf error.

Regards

[latifa]

Hi Erman,

I tried to delete the cloned server :

SQL> SELECT * FROM ADOP_VALID_NODES;

NODE_NAME       CONTEXT_NAME                   ZD_EDITION_NAME
--------------- ------------------------------ ------------------------------
udkfinp04       DBOFINP_udkfinp04              V_20170117_1619
udkfinp08       DBOFINP_udkfinp08              V_20170117_1619
udkfinp06       DBOFINP_udkfinp06              V_20170117_1619
udkfinp07       DBOFINP_udkfinp07              V_20170117_1619
udkfinp05       DBOFINP_udkfinp05              V_20170117_1619
udkfinp03       DBOFINP_udkfinp03              V_20170117_1619

with:

perl $AD_TOP/patch/115/bin/adProvisionEBS.pl ebs-delete-node \
-contextfile=

it doesn't work --> it's the master one.

What should I do at at this stage ?

Regards,
Latifa

[latifa]

Hi Erman,

I run: EXEC FND_CONC_CLONE.SETUP_CLEAN;

autoconf on DBTier and again my server --> it finished successfully.

I chekced ADOP_VALID_NODES. My server is registred with the good CONTEXT_NAME.

I will do preclone for the run file system and clone patch file system.

For the other severs may I use cloning with dualfs ?

Regards,
Latifa

[ErmanArslansOracleBlog]

Hi Latifa,

you didn't give all the details about what you are trying to do.
I think you are trying to clone a multi apps node (non-shared filesystem) EBS 12.2 instance.

Everyting about this operation is documented.
using dualfs is documented as well.

Steps for adding a node on a non-shared file system is also documented and using dualfs for that operation is also documented.

Ref: Cloning Oracle E-Business Suite Release 12.2 with Rapid Clone (Doc ID 1383621.1)

However, you are saying "I will do preclone for the run file system and clone patch file system"

So, what I understand from this is, you will not use dualfs for cloning one of your apps nodes and you are asking if it is ok to use dualfs for others.
My answer is yes. dualfs is recommended way of cloning anyways.