ORACLE APPLICATIONS ARCHIVES

Topicwise collection of
Postings on Mail Lists
ON
SYSTEM ADMINISTRATION



Purging Conc. Req. excluding specific Users

Date: Thu, 24 Jun 1999 12:09:48 -0400
From: "Bill Keenan" bkeenan@goodegg.com
Subject: Sysadmin Purge Concurrent Requests?

Hi All,

I would like to schedule my "Purge Concurrent Requests" to run daily, purging any requests over 12 hours old, however I would like to exclude one User from this request, as they have several reports which are too large too print, too large to export into excel, etc. Is there any way to exclude a user id???

If not, does anyone have any suggestions how I would purge the requests without purging a few specific requests from one user?

Thanks in advance for your help.
Bill


Date: Thu, 24 Jun 1999 13:43:11 -0500 (CDT)
From: "James J. Morrow" jmorrow@csac.com
Cc: jmorrow@csac.com
Subject: Re: Sysadmin Purge Concurrent Requests?

Unfortunately, the "Purge Concurrent Requests" process is written in the "positive" context, so it doesn't allow you to apply any negative logic (to exclude one person/process, etc).

But... the information required to purge the records is contained in the FND_CONCURRENT_REQUESTS table...

-- James
James J. Morrow E-Mail: jmorrow@csac.com
Systems Consultant USSnail: 6380 LBJ Freeway #181
Computer Systems Authority Dallas, Texas 75240
Dallas, Texas (972) 960-0180 x244
Digital Pager: (800) 759-8352 Pin# 1268587


Date: Thu, 24 Jun 1999 21:40:56 +0100
From: "Jonathan Stuart" jdstuart@globalnet.co.uk
Subject: Re: Sysadmin Purge Concurrent Requests?

Bill,

Unfortunately you can't do a purge and specify exceptions to the purge, only inclusions. In your situation the only way to purge everything except this user's output is to setup multiple purges which will cover all other requests but not this one, but this would probably result in dozens of purge routines being required, which is not ideal! Instead of that I would setup a small number of purges to clear down any large or particularly frequent jobs and run that daily, then have a "catch all" request which clears down everything older than, say, 1 week. This last job will make sure all jobs eventually get purged but should give your user time to do something with their output.

I hope this helps.
Jonathan.



Load the fnd_user table

Date: Thu, 24 Jun 1999 12:18:11 -0700
From: sunanda sunanda@smartdb.com
Subject: HELP!!on fnd_user

My client is asking us to load the fnd_user table directly . The problem I have is with the columns encrypted_passwords which are are not null and I have to load the columns. Now I have to encrypt the password before I load the columns.

I tried to locate the package that can do this. I think it is FND_CRYPT_PKG and the function in this that should be used is ENCRYPT. Now what variables to send this function.

Can anyone help me on this!!

Thanks
Sunanda
--
Sunanda Parvathaneny SmartDB Corporation
tel: (650) 988-8996 x229 1121 San Antonio Rd.
fax: (650) 988-6833 Palo Alto, CA 94303
sunanda@smartdb.com http://www.smartdb.com


Date: Thu, 24 Jun 1999 13:04:58 -0700 (PDT)
From: Andy Schindler aschindl@yahoo.com
Subject: Re: HELP!!on fnd_user

Sunanda,

Instead of fooling around decrypting and encrypting passwords, I would bring all of the users over defaulting the two password fields to "PASSWORD". Then I would change the password for sysadmin to "password" from the Define User Form.

Finally, I would run an SQL script to set all imported rows to "password" with:

UPDATE FND_USER
SET ENCRYPTED_FOUNDATION_PASSWORD =
(SELECT ENCRYPTED_FOUNDATION_PASSWORD FROM FND_USER
WHERE USER_NAME='SYSADMIN'),
ENCRYPTED_USER_PASSWORD =
(SELECT ENCRYPTED_USER_PASSWORD FROM FND_USER
WHERE USER_NAME='SYSADMIN'),
PASSWORD_DATE = NULL
WHERE ENCRYPTED_USER_PASSWORD = 'PASSWORD';

When the users log in for the first time, their password is "PASSWORD" and they are forced to change it to what ever they want. This is a lower security risk and a whole lot faster.

I have done this on 10.6 and 10.7 to demonstrate a security risk.

Lastly, this is getting into some pretty sensitive areas. As such, I am writing you directly and request that you keep this knowledge quiet.

Good Luck!
Andy Schindler
NRC, Inc.



Conc. Mgr - error from arreci

Date: Sat, 26 Jun 1999 13:07:07 +0300
From: mohammed alshaarawi m_sharawy@tccbce.com
Subject: concurent manager completed with error

hi

I have this error with a single user

Completion Message : ARXREC:Error from arreci.

what does it mean?
thanx
shaarawi


Automation of Conc. Mgrs. on NT

From: =?iso-8859-1?q?T=20L?= tal_2008_1@yahoo.com
Subject: Automation of Concurrent Managers on NT
Date: Mon, 28 Jun 1999 20:26:45 +0100 (BST)

Hello,

We've installed a dev instance of Rel 11 on NT 4.0 (Service Pack 3) and have created the concurrent manager service. Does anyone have anything to automate the startup and shutdown of the managers and the concurrent manager service?

I'd really appreciate any help.
Thanks,
Tracey.


From: sean a seansd@yahoo.com
Subject: Re: Automation of Concurrent Managers on NT
Date: Mon, 28 Jun 1999 12:41:46 -0700 (PDT)

when you create the concurrent manager on NT, you have the option to make the service manual or autostart and the service to depend on. here's the example from the cmsrvadm command:
cmsrvadm add APPS manual OracleStartORCL

the above would add the service oracleconcmgrapps, depends on the service oraclestartorcl, and is a manual startup. change the manual parameter to auto to automate the concurrent manager service. Because of the dependency created, starting the oracleconcmgrapps service will also start the oraclestartorcl service and bring up the database if it's shutdown at the time.

hope this helps, Sean


From: "Richard du Feu" Richardd@care.ca
Subject: Re: Automation of Concurrent Managers on NT
Date: Mon, 28 Jun 1999 16:03 -0500

Tracey,

We have created a batch file in NT to do that for our nightly shutdown for backup purposes, using the NET STOP command.

We created a batch file to start and to stop the concurrent manager service and then the database.

Here is what we put in our file:

Net Stop OracleConcMgrTEST. This stops the Internal concurrent manager service, which will terminate the other managers.

We used the AT command to schedule the batch file to start and stop and different time.

Create another batch file to schedule the times with the AT command
at 00:20 /every:monday,tuesday,wednesday,thursday,friday
"d:\release11\database\care\dba\SHUT_CONC.bat" TEST

If you want to pass a parameter for your DB instance, such as above you have to add the parameter to the AT line at the end (as we did with TEST). In this case your concurrent manager shutdown would have the following insteaf of the line being hardcoded:

NET STOP Oracleconcmgr%1

Unfortunately, the NET STOP command is not full proof. From time to time it will fail and will have to be started manually. For the most part it has been working fine. The only thing I can suggest is that you make sure there is enough time between shutting down the concurrent manager and the DB. The concurrent manager takes several minutes to shut down properly, however when you issue the NET STOP command it says it hasn't shut down properly and stops, but it actually issues the CONCSUB.exe. This is the process that kills all of the concurrent manager processes, it just takes a few minutes, the NET STOP command just doesn't wait around.

By the way to start you would do the same but create another batch file with the command NET START

Hope this helps,
Richard



Add a new Look-up Type

From: "Mike Schipper" mike.schipper@asd.bfg.com
Subject: Sys Admin Help on Look-up Type!
Date: Mon, 28 Jun 1999 15:17:30 -0500

We've been looking for the form that allows you to enter a new Lookup Type Code on the SO_Lookup_Types table. The doco tells us to go to the Special Quick Codes form, but that doesn't appear to be correct. We're on 10.7 character based Oracle. Does anyone know the navigation path to the form used to add new Lookup Types?


Custom.pll - 'committ' issue

Date: Wed, 30 Jun 1999 17:06:52 -0000
From: Yeche yeche@unitask-inc.com
Subject: Custom.pll

Hello All

Could you help me in 2 issues
1. I am trying to validate the item description when creating item (like I don't want that in the description will be the sine * ) I check it at the custom.pll at trigger WHEN-VALIDATE_RECORD If there is * in the description item I put an error MSG and then RAISE Form_Trigger_Failure the code is :

form_name varchar2(30) := name_in('system.current_form');
block_name varchar2(30) := name_in('system.cursor_block');
begin
if event_name = 'WHEN-VALIDATE-RECORD' then
if (form_name = 'INVIDITM') then
if (block_name = 'MTL_SYSTEM_ITEMS') then
if name_in('MTL_SYSTEM_ITEMS.DESCRIPTION') like '%*%' then
fnd_message.set_string('Please change description');
fnd_message.error;
raise Form_trigger_failure;
end if;
end if;
end if;
end if;

What I am getting is that I can see the message but the record is committed Does anyone have a suggestion ??

2. I need to add item to one of the forms. Can I do it within the custom.pll or I have to customize the form???

Thank you a lot
Yechiel Tushia
Unitask


Date: 30 Jun 99 11:06:42 MDT
From: vv vv venvivek@usa.net
Subject: Re: [Custom.pll]

1. Instead of issueing a raise form_trigger_failure U can simply try to give "return;"

2. Where U have given Ur commit statement. If Ur commit statement is going to be in Ur form then definetely it will commit. even U raise form trigger failure. To avoid that U can user on-commit trigger to check ur validation on the block level of ur block where which u wanna check.

3. U may check giving APP_EXCEPTION.RAISE_EXCEPTION instead of raise form_trigger_failure

What do u mean here an Inventory item or a new field. If it is goin to be a new field U have to customize Ur form.

I think these suggestion may help U.

With Regards.
Vivek.
Tech Consultant.
QuestAm.
vivek.venkataraman@weac.com



FND_REQUEST.SUBMIT_REQUEST arguments passing

Date: Mon, 28 Jun 1999 18:01:11 -0700
From: Vijay.Nallapati@Notes.airtouch.com
Subject: FND_REQUEST.SUBMIT_REQUEST arguments passing

Hi All,

Here I have created a procedure in which I submit a standard concurrent program "Analyzer - Extract Balances" to get the request id with which I pass a dat file name as a parameter to another conc. program . There are 10 parameters (arguments) for this program. Some of these have value sets and some have default values. But I am passing all the 10 parameters some with nulls . As this one is not from forms ,I am not passing all the arguments,i.e, not the other 90(as total arguments allowed are 100). When I submit this PL/SQL procedure in SRS I get an error wrong number or types of arguments. I don't think I need to pass 100 arguments when we submit this this way. Any error handling method for this will be appreciated.

Could anybody know about this problem? Pl. respond ASAP.
Thanks in advance, Vijay


Date: 30 Jun 99 11:40:43 MDT
From: vv vv venvivek@usa.net
Subject: Re: [FND_REQUEST.SUBMIT_REQUEST arguments passing]

Hi.
I believe Ur getting a error

Wrong number of arugument line 1 column 7. Right !!!.

Its manadatory to add to out parameters when U call ur procedure through the concurrent program.

Parameters R =

1. errbuf out varchar2.
2. retcode out number.

Just add this to Ur PL/SQL procedure. All set. Ur concurrent will work fine.

Vivek.
Oracle Consultant
QuestAm.
vivek.venkataraman@weac.com


Conc. Mgr. - where to place

Date: Thu, 01 Jul 1999 09:56:37 -0500
From: Kevan Forest kevan.forest@uni.edu
Subject: Where to place concurrent managers?

Hi,

We are currently working to define how we plan to install the Oracle Apps in our environment and have little first hand experience.

We are currently planning to place the DBMS server and the Apps server on separate sun boxes. Sun in their efforts to help us size the boxes would like to know where we plan to have the concurrent managers run. My first thoughts on this were to put them on the applications server, but in reviewing old postings to this list I see many are running them on the DBMS server.

My question is, what are most other shops that have separated the Apps and DBMS servers, doing with their concurrent managers and why?

Thanks!
Kevan Forest
University of Northern Iowa


Date: Fri, 2 Jul 1999 09:18:33 +1000
From: "Bevin Watson" bevin.watson@abs.gov.au
Subject: Re: Where to place concurrent managers?

I'm not particularly expert on this, but I would place the CMs on the DB Server because:

1. CM jobs attack the database, so they should be as close as possible to the database.
2. You will always have 1 database server, you may have multiple application servers.
3. The patches come with the reports in the "database server" patch set - so Oracle are assuming the CMs are on the database server. If you are very big, talk to Oracle. They have a special team who love to play with big databases.

Bevin Watson
Financial Systems Administrator
Australian Bureau of Statistics


Date: Fri, 2 Jul 1999 12:22:58 -0500
From: "Fischer, Craig J." FischerCJ@bvsg.com
Subject: RE: Where to place concurrent managers?

In general, I believe Oracle's recommendation is that they be placed on the database server. We have a three-tier configuration running 11.0.3 on AIX but are not following this standard.

For one, we only have 1 middle-tier server at the moment; at whatever point we need to add another one, you can bet we'll move the concurrent managers to the DB server (so that they don't have to be maintained on two or more servers). Also, our two servers are in a cabinet and connected via a high-performance switch; this provides near backplane speeds and is much faster than Ethernet. If we could not take advantage of this connection, we would have seriously considered moving the concurrent managers to the DB server.

The main reason we like our current setup is that Oracle Apps patches only need to be applied on one server, the middle tier. Since we have only the database on the DB server, Apps administration is somewhat simpler than it would otherwise be. In addition, it helps separate purely application-related code from purely database-related code. We also have enough memory to run the CM's on the application server.

One thing we have to be sure of, though, is that we keep the database installations on both servers at exactly the same patch level. But since we don't patch the database nearly as often as we patch the Apps, this isn't a big deal.

Craig Fischer
BV Solutions Group
fischercj@bvsg.com


Date: Fri, 2 Jul 1999 12:58:58 -0500
From: Bharat Patel bpatel@DOMINOAMJET.com
Subject: RE: Where to place concurrent managers?

we have it on application server. we are running with one db server and one apps server. we have 2GB RAM on Apps server as well as DB server with 3 CPU on each one.

Bharat Patel
D.B.A.
bpatel@dominoamjet.com
847-244-2501 ex 1249



Apps going into a hang

Date: Thu, 01 Jul 1999 11:26:55 -0700
From: Cyndie Winrow cwinrow@mail.sdsu.edu
Subject: running JDK 1.1.5.18o w/ r11.0.2

Hi All ~ We're running apps 11.0.2 on Sun Solaris 2.6, database 8.0.4x; in addition most of our users are Win95 (minority using WinNT) and have the Novell client installed and active on their workstations...

Find that often they "freeze" in the apps during the afternoon - cannot navigate out of the form, close the form, etc. But they can still use other programs on their desktop (internet, email, etc.). Have wondered if there is a memory management issue for the desktop since it appears the applet invokes a DOS session...

Support gave me two commands to append inbetween the appletviewer.exe and the server url, one dealing with memory and the other dealing with caching... Has anyone experienced this same apparent apps freeze problem and have you tried these settings? Also, has anyone upgraded from the 1.1.5.18o applet to the 1.1.5.21o applet and found better performance? Problems?

Thanks in advance ~
Cyndie Winrow
Oracle System Administration
San Diego State University
619.594.0864


Date: Thu, 1 Jul 1999 11:38:53 -0700
From: skirby@bco.com
Subject: RE: running JDK 1.1.5.18o w/ r11.0.2

Do you also get FRM-99999 Errors?

The problem we have seen this with is when the network traffic to too high and packets are getting dropped. It takes Oracle a while to catch a dropped packet and during that time it is frozen. If it doesn't pick up the packet it will crash with a FRM-99999.

If I were a betting man I would guess that this is occurring between 1-3pm ... That is when our network has its peak time. Everyone feeling guilty for not getting anything done in the morning and checking e-mail after lunch...

Steve Kirby
skirby@bco.com


Date: Thu, 1 Jul 1999 12:21:27 -0700
From: Andy Baliton abaliton@us.salterlabs.com
Subject: RE: running JDK 1.1.5.18o w/ r11.0.2

We have the same issue in here using apps 11.0.2 sun solaris 2.6 database 8.0.5. We have upgraded our forms to 4.5.10.10.2 and jdk to 1.1.21o. It helps a lot. Sometime, if 6 or more concurrent manager are submitted, it will slow down the system and causes some workstations to freeze. We are in the process of testing Netscape or Internet Explorer to access our apps hopefully this it will help.



CUSTOM.pll: WHEN-VALIDATE-RECORD event

Date: Thu, 1 Jul 1999 12:58:47 -0400
From: "Annamaneni, Hanumantha" Hanumantha.Annamaneni@alcoa.com
Subject: CUSTOM.pll: WHEN-VALIDATE-RECORD event

I am trying to use the WHEN-VALIDATE-RECORD event in CUSTOM.event procedure for three forms - one INVTTMTX, WIPTXCMP and WIPTXMAT forms. Using some debug statements I realized that the event WHEN-VALIDATE-RECORD is passed to the procedure for only the INVTTMTX and not for the WIP forms. The documentation suggests that every form should send the WHEN-VALIDATE-RECORD event to the custom.event procedure. Has anybody come across this situation before? Any feedback would be appreciated.

Thanks Hanumanth


Date: Wed, 14 Jul 1999 16:05:36 -0000
From: Yeche yeche@unitask-inc.com
To: "'oraapps-l@cpa.qc.ca'" oraapps-l@cpa.qc.ca
Subject: RE: CUSTOM.pll: WHEN-VALIDATE-RECORD event

Hello
There is patch that fix the raise_Form_Trigger_Failure The patch number is 523460

Hope it will help Yechiel Tushia
Unitask-Consulting


Conc. Prog. - Parameters clash

Date: Fri, 2 Jul 1999 09:42:39 +0100
From: rchalton@dunnes-stores.ie
Subject: Concurrent Program Parameters and Validation Sets

I have a concurrent program that requires three parameters, say P1, P2 and P3. All are optional, but if specified the following rules apply:

P1 must be a date in an Open or Future Enterable period, and must be specified alone, i.e. P2 and P3 not allowed.

P2 must exist in table T1 column C1 and requires entry of P3, but P1 not allowed.

P3 must be a date in an Open or Future Enterable period, and requires entry of P2, but P1 not allowed.

It is not an option to specify as two concurrent programs, the first with parameter P1 only and the second with mandatory parameters P2 and P3 only.

I currently handle this validation at the concurrent program code level, but would rather handle it at the parameter level.

Any suggestions would be greatly appreciated.



Setting Multiple Userids to Employee

Date: Fri, 2 Jul 1999 11:32:19 -0400 (Eastern Daylight Time)
From: Mary Ann Carr macarr@cmu.edu
Subject: Setting Userids to Employee

Oracle System Administration permits setting multiple user ids to the same employees but it gives a message that says "you may cause problems in the application" by doing so. We plan on doing this to facilitate training exercises.

Does anyone know what the potential problems are?
Thanks in advance!
----------------------
Mary Ann Carr
Financial Management Project
mc4t@andrew.cmu.edu


Date: Fri, 2 Jul 1999 10:45:03 -0500
From: "Logan, Ernie" Ernie_Logan@bmc.com
Subject: RE: Setting Userids to Employee

I could see it causing problems with SSHR (Self Service Human Resources) if two user IDs associated with the same employee attempted to perform simultaneous updates. Other than that, it could also possibly cause problems with workflow notifications.


Date: Fri, 2 Jul 1999 11:48:31 -0400
From: "Devulapally, Sridhar, CFCTR" sdevul0458@att.com
Cc: macarr@cmu.edu
Subject: RE: Setting Userids to Employee

Hi Mary

I have done this and am not aware of any problems. Everything is running fine for us. The only reason we had to assign all the users of AP to an employee is to View the POs created by Employees. I have assigned the same employee to about 600 users and no problems since an year.

Hope this helps
Thanks
Regards
Sridhar Reddy
(407) 858-7610
sdevul0458@att.com mailto:sdevul0458@att.com


Date: Fri, 02 Jul 1999 11:42:46 -0400
From: Jennifer Knice JKnice@cerbpyro.com
Subject: RE: Setting Userids to Employee

One potential problem involves approval routings (including purchase orders, requisitions, etc.). Documents get routed based on the employee name. If multiple sign-ons are linked to that employee, only one individual, based on standard functionality, will get notified and can reply.

Jennifer Knice


Date: Fri, 2 Jul 1999 12:43:22 -0400
From: Carolyn DeSimone cdesimone@goaa.org
Subject: RE: Setting Userids to Employee

You may run into problems if you are training in the Purchasing module. For instance, if the same employee is searching the same LOV at the same time, the system will slow down. Depending on how may training instances you are setting up, I find it easier to assign a different employee to each Student ID, it gives variety to the results. We are using a hierarchy to route requisitions and releases through the system, so I have to be careful to choose employees that are part of the hierarchy and have approval authority.

Carolyn DeSimone
Purchasing Oracle Implementation
Cdesimone@goaa.org
407-825-2028


FND_FILE - ORA-600 error

Date: Fri, 2 Jul 1999 17:13:06 +0100
From: rchalton@dunnes-stores.ie
Subject: FND_FILE

I'm getting ORA-600 when trying to use FND_FILE.PUT_LINE in a packaged procedure called from a concurrent program (SQL*Plus). We are running Financials Release 11.0.2 on Windows NT. Any ideas greatly appreciated.


Date: Fri, 2 Jul 1999 10:27:04 -0600
From: PKBATCHU@micronpc.com
Subject: RE: FND_FILE

FND_FILE is package is used in forms. You may need to use 'UTL_FILE' in PL/SQL.


Date: Fri, 2 Jul 1999 12:07:40 -0500
From: "Logan, Ernie" Ernie_Logan@bmc.com
Subject: RE: FND_FILE

This is simply not correct. FND_FILE is specifically for use in PL/SQL (concurrent programs only). PKBATCHU is thinking of TEXT_IO.

I can't help much with the ORA-600 (it's a generic 'internal database' error), except to suggest that you have your DBA set a trap for this in init.ora, and do a trace dump. Contact WWS for analysis. What RDBMS version are you running? There are quite a number of documented ORA-600 problems fixed in the 8.0.5.1.0 patchset.

FND_FILE is very dependent on the utl_file_dir specified in init.ora. The oracle user account must have write privileges to this directory. Even though applmgr is the user of the concurrent manager, FND_FILE uses UTL_FILE in the RDBMS to write the output. This same directory must also be set as APPLPTMP. Please verify these settings. Failure to set one of these could be the cause o your problem.


Date: Fri, 2 Jul 1999 13:01:35 -0500
From: Bharat Patel bpatel@DOMINOAMJET.com
Subject: RE: FND_FILE

which version of database you are? if it 8.0.4.2.1 then this is a known bug.

Bharat Patel
D.B.A.
bpatel@dominoamjet.com
847-244-2501 ex 1249