Wednesday, June 27, 2007

Multi to Single node Cloning issue

A friend called me yesterday he was facing the following error when he tried to clone from multi node to single node cloning
"Could not create the high water mark manifest..Please re-run the program to create the manifest required for merging APPL_TOPs"

He faced this error after he run the (perl adpreclone.pl appsTier merge) and then he cancel it while running and then he run it again on the other node then he recived this error.

According to metalink notes 372819.1 and 372758.1 he should reapply the AD.I.x patch again to the node and then run Maintain snapshot and then run the perl adpreclone.pl appsTier merge again, we did so and every thing went fine.

Hope that helped and by the way this will be my last post before my Eye LASIK Surgery which is scheduled tomorrow morning at 8:30 a.m.

Fadi

Sunday, June 24, 2007

Support forum,

I am thinking of opening a forum here to share support cases with all users, even though I guess oracle forums a great place to post problems and to get answers quickly, I will try to open my own forum for 3 months period if it was successful I will keep it other wise I will stick with oracle forums.
Suggestion are welcomed
Fadi

Tuesday, June 19, 2007

Oracle support day in amman

Been busy lately, yesterday I went to oracle support day at Hayyat Amman. It was presented by two oracle support managers, wael sadaqa from Oracle dubi and jafar al omary from Oracle Saudi Arabia, they presented how to work effectively with support and the escalation process and they also stressed on using oracle diagnostics and configuration support manager. They mentioned a little bit about R12 upgrade process and they mentioned that people who are doing upgrade will get more attention from oracle support.
It was good and helpful, the presentations can be downloaded from here
Fadi

An update on this (25-jun-2007)
The presentations can be downloaded from here

Monday, June 11, 2007

Open Workflow notifications list is very slow

Here I am publishing a note which I already published on metlaink through the Customer Knowledge Exchange program the note id is 428573.1, they had this program a couple of months ago and I managed to publish two notes and I will be sharing them here with you.

I faced a slow performance issue which only applies on the notifications list page, opening the workflow notifications list to see a list with two notifications when the problem happened used to take a lot of time (more that 20 minutes and most of the times it used to time out) so I followed the below to solve the issue

SQL> select message_type,count(*) from WF_NOTIFICATIONS group by message_type;

MESSAGE_ COUNT(*)
-------- ----------
AZNM000 3
FERASWF 8
POAPPRV 121
POERROR 29
PORCPT 38
POSDSALE 332
REQAPPRV 131
WFERROR 1137436

8 rows selected.
SQL> select count(*) from WF_NOTIFICATIONS where status='OPEN'
and MESSAGE_TYPE='WFERROR'

998976 rows selected.

I find out that I have a lot of the WFERROR message_type and most of them with an open status so I took a backup of the table just incase any thing went wrong using the following statement

SQL> CREATE TABLE WF_NOTIFICATIONS_bck AS SELECT * from WF_NOTIFICATIONS;

Table created.

Then I run the following update statement to set the status of all open notifications of the WFERROR type

SQL> update wf_notifications
set status='CLOSED', mail_status='CANCELED', end_date=sysdate
where status='OPEN'
and MESSAGE_TYPE='WFERROR' ;

998976 rows created
SQL> commit;

After that I tried the notifications list page it did not take more than 3 seconds and every thing went back to normal, so I scheduled a database job that run every month to close all the open notifications with message_type wferror.

To avoid this problem from happing again I am make sure that all the sysadmin notifications are closed and I also have monthly task to check that the above job is running.

i hope that helped
fadi

Friday, June 08, 2007

Special thanks to some of Oracle teams

I would like to take the opportunity to thank OCP exam teams, both teams (ocpreq and ocpexam) showed high level of support and fast response time on my inquiries regarding the 11i Apps DBA OCP, they supported me step-by-step to submit my hands-on-course.

In the same time I would like to give a special thanks to oracle Customer Knowledge Exchange team a.k.a CKE team for their support and help when I tried to submit an article to be published at metalink, they showed a high level of support. a big thanks goes to (Mary Anne,Debbie Kiaaina) from this team.

So thank you ocp team/cke team and keep up the good work
Fadi

Monday, June 04, 2007

Concurrent request and its database SID

I had a long running Concurrent request and I needed to monitor it from the Enterprise manager but I needed its Database SID, so I search on metalink and I found this query which helped me out.

Connect as apps user

SQL> column process heading "FNDLIBR PID"
SELECT a.request_id, d.sid, d.serial# ,d.osuser,d.process , c.SPID
FROM apps.fnd_concurrent_requests a,
apps.fnd_concurrent_processes b,
v$process c,
v$session d
WHERE a.controlling_manager = b.concurrent_process_id
AND c.pid = b.oracle_process_id
AND b.session_id=d.audsid
AND a.request_id = &Request_ID
AND a.phase_code = 'R';


You need your concurrent request ID as an input.
c.SPID= is the operating system process id
d.sid= is the Oracle process id

I hope it will help you


Fadi

Saturday, June 02, 2007

Workflow again

-an update on this we applied the patch and it solved our issue
Some end-users open cases with the support claiming that they don’t receive all the email notifications related to some of their purchase orders, I thought that maybe their email were spelled wrong or maybe they changed their email preference so they don’t receive an emails at oracle application level like the last time, I checked both but none of the above applies on the case so I had to check an one of there POs on the wf_notifications table, I run the following query

select * from wf_notifications where subject like '%200700001052%'

and the email_status value was ERROR so I took the notification_id and search for it in the emails that were sent by the notification mailer to the sysadmin and I found an email with the following content

"Event Error Name: -20001Event Error Message: ORA-20001: Oracle error -20001: ORA-20001: APP-FND-02500: Error occurred during product initialization for MO when executing 'begin MO_GLOBAL.INIT; end;'.SQLCODE = -20001 SQLERROR = ORA-20001: APP-FND-02902: Multi-Org profile option is required. Please set either MO: Security Profile or MO: Operating Unit profile option.has been detected in FND_GLOBAL.INITIALIZE."

I check that error on metalink and I found 5 notes saying that I must apply patch (5209533) p5209533_11i_GENERIC.zip

I will apply that patch on my test/clone instance this weekend and I will update you if the issue was resolved.

For more info you can check note (393847.1) or search for the error/patch


Hope that helped
Fadi