Sunday, July 3, 2016

Blocks corrupted in system tablespace after fresh installation of EBS 12.1.1

Today, i encountered an issue where i installed the EBS 12.1.1 and tried to take a full rman backup.
First converted the db to archive mode and then tried to take a full rman backup.
Then configured the rman channel to place the backup in disk at the designated location.

RMAN backup error.

thankfully, these corruptions does not belong to any object and hence it will not impact the installtion or db startup.

To verify the blocks corrupted, we need to run the dbv command line utility:
[oracle@sam fmw]$ dbv file=/u01/12.1.1/EBS/SOURCE/db/apps_st/data/system09.dbf

DBVERIFY - Verification complete

Total Pages Examined         : 192128
Total Pages Processed (Data) : 52304
Total Pages Failing   (Data) : 0
Total Pages Processed (Index): 26450
Total Pages Failing   (Index): 0
Total Pages Processed (Other): 2123
Total Pages Processed (Seg)  : 0
Total Pages Failing   (Seg)  : 0
Total Pages Empty            : 84778
Total Pages Marked Corrupt   : 26473
Total Pages Influx           : 0
Total Pages Encrypted        : 0

Highest block SCN            : 3712725850 (1388.3712725850)





Once you conform through dbv the blocks are corrupted then you need to run the Rman validating script 

Rman script to  validate

run {
allocate channel d1 type disk;
allocate channel d2 type disk;
allocate channel d3 type disk;
backup check logical validate database;
release channel d1;
release channel d2;
release channel d3;
}

it  will produce the error in the alert log for all the corrupted blocks it will validate all the blocks corrupted and populate into the v$database_block_corruption Once the Rman command completes then you can proceed to the next step ..it is mandatory to complete this step



Check the no of blocks corrupted

SQL> Select * from v$database_block_corruption ;

     FILE#     BLOCK#     BLOCKS CORRUPTION_CHANGE# CORRUPTI
---------- ---------- ---------- ------------------ ---------
       352     165660      26341                  0 ALL ZERO

note the  it is the data file number which is effected  and the block number



Check the corrupted block belongs to any object


 SYNTEX :select segment_name, segment_type, owner
       from dba_extents
      where file_id = <data_file_name>
        and  < corrupted block number >between block_id
            and block_id + blocks -1;



SQL> select segment_name, segment_type, owner
       from dba_extents
      where file_id = 352
        and  165660 between block_id
            and block_id + blocks -1;

It will return now rows selected and then proceed with the steps if it returns now row selected then the particular block  does not belongs to any object and it is empty
To conform the empty block run the below query



Then, i performed the below action plan

Connected to:
Oracle Database 11g Enterprise Edition Release 11.1.0.7.0 - 64bit Production
With the Partitioning, OLAP, Data Mining and Real Application Testing options

SQL> alter database datafile '/u01/12.1.1/EBS/SOURCE/db/apps_st/data/system09.dbf' resize 1056M;

Database altered.


Verified again the count of corrupted objects

[oracle@sam 11.1.0]$ dbv file=/u01/12.1.1/EBS/SOURCE/db/apps_st/data/system09.dbf

DBVERIFY: Release 11.1.0.7.0 - Production on Sun Jul 3 10:54:10 2016

Copyright (c) 1982, 2007, Oracle.  All rights reserved.

DBVERIFY - Verification starting : FILE = /u01/12.1.1/EBS/SOURCE/db/apps_st/data/system09.dbf


DBVERIFY - Verification complete

Total Pages Examined         : 135168
Total Pages Processed (Data) : 104090
Total Pages Failing   (Data) : 0
Total Pages Processed (Index): 26450
Total Pages Failing   (Index): 0
Total Pages Processed (Other): 2124
Total Pages Processed (Seg)  : 0
Total Pages Failing   (Seg)  : 0
Total Pages Empty            : 2504
Total Pages Marked Corrupt   : 0
Total Pages Influx           : 0
Total Pages Encrypted        : 0
Highest block SCN            : 3714312362 (1388.3714312362)
[oracle@sam 11.1.0]$ 


Re initiated the backup again and it completed successfully.


Reference
=======
1. Oracle E-Business Suite Installation and Upgrade Notes Release 12 (12.1.1) for Linux x86-64 [ID 761566.1]
2. More than 20000 Blocks Corrupted in system09.dbf After Fresh Installation of E-Business Suite R12.1.1 on Linux x86_64 Platform [ID 840963.1] 



Sunday, June 12, 2016

Analyze patches using Patch Wizard

Before applying any patch it is very important to know the prerequisites. We can get the information from read me and also through Oracle provided Patching Wizard.

In my system i have not provided the metalink credentials to patch wizard, thus i'll be downloading the patches manually and then place it inside the staging area in the server.

I specified my stage directory as /home/oaa/appldev/patchanalysis and the patch to be analyzed is-21980909. The stage directory is set by the option-Patch Wizard Preferences.



When i ran it i got the below error:

Patchlist: 21980909 Platform: 226 Languages:

Validating Staging directory '/home/oaa/appldev/patchanalysis'
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Warning: Your MetaLink Credentials are not set. Download of patches skipped.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
End of Patch Download Program for patch list.... [Jun 6, 2016 2:19:13 AM]

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Calling Patch Upload Program for Patchlist.... [Jun 6, 2016 2:19:13 AM]
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Starting Upload Process...
 Parameter List  21980909

Program Execution Start time: [Jun 6, 2016 2:19:14 AM]

Stage Dir is  /home/oaa/appldev/patchanalysis
Entity hastable created. Total entries : 239

Starting processing patch 21980909 at [Jun 6, 2016 2:19:14 AM]
ERROR: No patch zip file was present under stage directory

Program Execution End time: [Jun 6, 2016 2:19:14 AM]
===================


From 12.1 onwards, we have to specify ad/ nonad under staging directory. I set the new staging directory as /home/oaa/appldev/patchanalysis/ad and rerun the patch wizard(Recommend/Analyze patches option.

References:https://docs.oracle.com/cd/E18727_01/doc.121/e12148/T531058T531063.htm

The program completed with warning because the metalink credentials were not provided.





Click on Sub requests for more details




Click on details option. It will show the patch description, the prerequisites etc.







                                                                                                                                                  





























Friday, June 3, 2016

EBS 12.2.5 Demo image Compute Cloud

Whoever want to have a hands on EBS 12.2.5 cloud, he/she can refer the below note.

https://cloud.oracle.com/marketplace/listing/5248518?_afrLoop=4737445521078117&_afrWindowMode=0&_afrWindowId=null--for image copy

Doc id-Doc ID 2066260.1

Community discussion-https://community.oracle.com/thread/3932643?sr=stream&ru=2793050


Wednesday, June 1, 2016

After clone, not able to see log and out file from front end

Today we had an issue where just after recent refresh, users started complaining that they are not able to see the concurrent program log and out files. The error was "The Applications File Server could not open the file "path" for read.

How i did the troubleshooting.

1. We all know that RRA(FNDFS executable) is the default text viewer to view the out and log files. I verified that the profile option "RRA:Enabled is set to yes or not. It was set to yes in my case. If not, then change it to yes.

2. Then i verified whether the apps listener was running in all the nodes(it was shared appl top with 2 apps tier and 2 cm tier). It was running in all the nodes.

3. Then i started checking the files inside $TNS_ADMIN because RRA uses configuration file tnsnames.ora and listener.ora. As it was shared appl top so the FNDFS and FNDSM entry should have the details for all the nodes in tnsnames.ora file. In my case, there were few entries missing and i had to add them manually.

tnsnames.ora file.

example-FNDFS_nodename=(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=NODENAME)(PORT=1234))
(CONNECT_DATA=(SID=FNDFS)))

FNDFS_nodename_domainname=

FNDFS_SID_nodename=

Repeat the same and add all the nodes like above.

In the same way, add the FNDSM entries.

And then create a ifile and add all the entries.

ifile entry

FNDFS_APPLTOP_nodename1=(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=NODENAME1)(PORT=1234))
(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=NODENAME2)(PORT=1234))
(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=NODENAME3)(PORT=1234))
(DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=NODENAME4)(PORT=1234))
)
(CONNECT_DATA=(SID=FNDFS)))

Note: The FNDFS_APPL_TOP_nodename1 should have CM1 as first entry in app1 and app2 server.

Repeated the same in all the nodes and bounced the ebs services. Asked the users to validate the same and they confirmed that it is working fine.


Sample TNSNAMES.ORA(END Portion)

FNDFS_APPLTOP_nodename1=
        (DESCRIPTION=
            (ADDRESS_LIST=
            (ADDRESS=(PROTOCOL=tcp)(HOST=CM node hostname)(PORT=port no))
                (ADDRESS=(PROTOCOL=tcp)(HOST=app2 node)(PORT=port no))
                (ADDRESS=(PROTOCOL=tcp)(HOST=app1 node)(PORT=port no))
                (ADDRESS=(PROTOCOL=tcp)(HOST=cm 2 node)(PORT=port no))
                (ADDRESS=(PROTOCOL=tcp)(HOST=dmz node)(PORT=port no))
                (ADDRESS=(PROTOCOL=tcp)(HOST=dmz node)(PORT=port no)))
            (CONNECT_DATA=
                (SID=FNDFS)
            )
        )









Friday, May 27, 2016

Running Oracle EBS in Google Chrome

Just to make everyone aware, Google Chrome is not not yet certified for running EBS forms. There are few specific versions for chrome which are only certified to run web based content with EBS but not java based forms.

The certification from Oracle covers(OA Framework/HTML). But there are always a workaround for every case(Mind it if any issue occurs you cannot log any request with Oracle).

There is an extension available within Chrome 48+(https://chrome.google.com/webstore/detail/thinforms-plugin-for-orac/nmjefghbgfcpoobigfbalocpncklkjhk?hl=en). This extension can be used to run java based content from EBS.

As per Steven Chan's blog https://blogs.oracle.com/stevenChan/entry/jre_ebs_directions
next EBS version is expected to use Oracle FORMS 12c, which will utilize java web start technology, which means it will be browser independent to use to launch the application.




Wednesday, May 25, 2016

Punchout issue in EBS 12.1.3

Today we encountered an issue where some users were not able to do a punchout to the supplier. When supplier was contacted they confirmed that they have renewed their certificates few days back.
Thus we figured it out that ssl handhake between EBS and supplier site is not happening and we have to import the renewed certificate to our Apache certificate.


When a punchout is attempted to a supplier, using https, in order to create a Secure connection between the Oracle E-Business Suite Middle Tier Server and the Punchout Destination, the certificate must be present in the certificate file that is used by the E-Business Suite Middle Tier - Apache Web Server.

Below are the steps.

1. Get the punchout url. Login into EBS>iprocurement responsibility>home page>stores>current stores>identify the supplier >and click on update>punchout definition>copy punchout url.

2. Open the url in IE, choose the certificate and export it in X-509 Certificate with chain (PEM)

3.  Save it in desktop

4. Now go to System Administrator - and move to  Profile/System - to query on the Profile - POR: CA      Certificate Filename. We will get value as $ORACLE_HOME/sysman/config/b64InternetCertificate.txt

5. Take a back up of this file and add the content of saved certificate to this file.

6. Bounced Apache and verified that the ebs is able to do a handshake now.

Transparent Data Encryption related docs

http://www.oracle.com/technetwork/database/security/tde-faq-093689.html#A13018

http://docs.oracle.com/cd/B19306_01/network.102/b14268/asotrans.htm#CHDCCICE

http://docs.oracle.com/cd/B19306_01/network.102/b14268/asotrans.htm#BABEBFBA