Blog dedicated to Oracle Applications (E-Business Suite) Technology; covers Apps Architecture, Administration and third party bolt-ons to Apps

Tuesday, October 23, 2007

FNDFS_ and FNDSM _ in tnsnames.ora of 806 home

The tnsnames.ora file of the 806 home has many entries starting with FNDFS_ and FNDSM_.

FNDFS or the Report Review Agent (RRA) is the default text viewer within Oracle Applications, which allows users to view report output and log files. Report Review Agent is also referred to by the executable FNDFS. The default viewer must be configured correctly before external editors or browsers are used for viewing requests. Check out metalink note 111383.1 which describes the Basics About Report Review Agent (FNDFS) on 11i.

FNDSM is the Service manager. Sometimes after cloning you may find that concurrent managers do not start up. Metalink Note 305986.1 describes the scenario:

Symptoms

You have followed instructions within Note 230672.1 - Cloning Oracle Applications Release 11i with Rapid Clone and completed the cloning successfully.

However, now the Concurrent Manager cannot be started. The following error is produced in the Internal Concurrent Manager log file:

Could not start Service Manager FNDSM_hostname.
The TNS alias could not be located, the listener process on hostname could not be contacted, or the listener failed to spawn the Service Manager process.

Cause

This is caused because FNDSM_hostname service is not present in the target instance.

FNDSM database trigger is used to create the FNDIM_ and FNDSM_ concurrent managers. This trigger is present on FND_NODE table. If triggers are disabled in source then issue may occur.
Solution

- Run this following select SQL statements on the target node only to check trigger status:
> SELECT trigger_name , status
FROM user_triggers
WHERE table_name = 'FND_NODES' ;

- If status of the triggers show as DISABLED, then enable these following triggers as follows:
> Connect apps/apps
> alter trigger UPNAME enable;
> alter trigger FNDSM enable;

- Take backup of FND_NODES table in Target.
> create table FND_NODE_bk as select * from FND_NODES ;

- Login into Oracle Application with System Administrator Responsibility.
- Delete the rows from FND_NODES table in target using install -> node -> Define form.
- Then re-Register TRAGET hostname.
- This should register FNDIM_ and FNDSM_ concurrent managers in FND_CONCURRENT_QUEUE Table.

For cluster configuration, when the cluster name is manually inserted in FND_NODES, you can do the following to ensure that FNDFS_ entries are created in tnsnames.ora :

From the Application tier:

1. Log in as applmgr
2. cd to $FND_TOP/patch/115/sql
3. Run the script: afdcm037.sql
4. This script will create libraries for FNDSM and create Managers for Preexisting Nodes.

Note: Service Manager "FNDSM" can not be created from form:
Concurrent> Manager> Define under Sysadmin Responsibility.

Reference Bug 6085070 FNDSM TRIGGER CAUSES SERVICE MANAGER NOT TO BE
CREATED AFTER CLONING SINGLE NODE

11 comments:

Sandeep Sulakhe said...

You have written FNDSM as Standard Manager initially. Its got to be the Service manager

Vikram Das said...

You are absolutely right Sandeep. I have corrected it in the post. Thanks for pointing it out.

- Vikram

Anonymous said...

Fantastic. Your FNDSM creation script hint was superb. I thank you a lot for putting it here.

thanks sir.

maganti said...

Hi,
Thanks for sharing the information about the FNDFS and CM's. I am new to oracle applications DBA. Here why we are deleting the data and how to register target hostname ?

Vikram Das said...

Hi Maganti,

After every clone, when adcfgclone.pl is run it cleans up the fnd_nodes table and repopulates it with the correct hostnames of the target instance during autoconfig. Sometimes this doesn't happen and the fnd_nodes table still has the old data. That is why if you follow what is recommended in this article, your issue may get resolved. The other way of getting around this is given in my post: http://oracleappstechnology.blogspot.com/2007/09/good-post-clone-procedure.html

- Vikram

harsha said...

Hi Vikram,

Its really a good documentation and helped me a lot to know about FNDSM.....
And also i have a doubt regarding the FNDSM.

In my production instance - It shows "Service-Manager: ERPTEST"

ERPTEST - is the Test server Hostname.
And for production Hostname is SALERPAPP.
Please let me know how i can move with this.

Thanks,
Harsha

Anonymous said...

Vikram,

I had an issue on a cloned 11i Windows environment, where TNSLSNR80.exe failed to spawn FNDSM.exe (every time, the process started and stopped).
Thanks to your article I found out that there was no "FNDSM_" entry in fnd_concurrent_queues. Running script afdcm037.sql was the solution.

Thank you very much!
Bernard

amit said...

Hi Vikram,

I was googling for one of my issue and reached to this older post.

Issue: After cloning I could still see the tnsping process to my old env

$ ps -ef |grep -i auxpdb02
appeoltp 4559 4097 1 06:45:44 pts/0 0:00 grep -i auxpdb02
appeoltp 4557 8197 2 06:45:43 ? 0:00 tnsping FNDSM_AUXPDB02_zeoltp
$

AUXPDB02 is my PROD server (Source)
but I could see that FNDLIBR (cpmgr) process invokes this child process every 30 seconds.
I have verified all my files (Listner.ora, Tnsnames.ora, Env files, Context_file) and there are no entries for AUXPDB02. Also checked the FND tables as well.
But still the process is pinging.

Need to know how this tnsping process is invoked and from where it is picking the refernce to PROD server.

Thanks for your valuable posts. I am regualr reader and a very huge fan of your knowledge sharing with all.

Regards,
Amit Garg

Martin said...

Hi,

i have the same problem of Amit.

After clone the target is still looking for the source (FNDSM) when i start the concurrent manager even if my config (files and tables) is ok.

Do you have an answer for me ??

Martin

nunna sindhu said...

Hi,
I was facing a similar problem while bringing up concurrent managers in oracle applications 11i.
Error is:
Could not contact Service Manager FNDSM_CRMAPP-UAT-2_EBCRMUTN. The TNS alias could not be located, the listener process on CRMAPP-UAT-2 could not be contacted, or the listener failed to spawn the Service Manager process.
Could anybody suggest on this.Thanks in advance.

Sindhu Nunna

Siddhartha Mukherjee said...

Hi All ,
For skiping this kind of issues just take a backup of listener.ora and tnsnames.ora file before clone and revert back after the clone.

In this scenario , first shutdown the applications 9kill all the process). Check if any FNDLIBR process is there or not. If not then run cmclean.sql first and do commit.
then run script afdcm037.sql ($FND_TOP/patch/115/sql to register the service manager and then try to up the ICM using scripts adcmctl.sh start. and check the manager log. Hopefully issue will be resolve :)