Invalid Objects
Date: Tue, 20 Jul 1999 14:38:33 +0100
From: jem.lough@blackwell.co.uk
Cc: James.Walker@blackwell.co.uk
Subject: Invalid Objects
Hi you Apps gurus!
We are experiencing a number of invalid objects appearing from time to time.
These cause various errors in the Client Windows Applications like :-
"
FRM-40505: ORACLE error: unable to perform query.
-----------------------------------------------------
ORA-04045: errors during recompilation/revalidation of
APPS.AR_PAYMENT_SCHEDULES_V
ORA-00600: internal error code, arguments: [17005], [3], [2399314652],
[2395565268], [2419283116], [], [], []
"
Running the query .....
SQL select owner, object_name, object_type from dba_objects where status =
'INVALID';
shows about 150 objects invalid.
These all go away when the invalid objects are recompiled via the script at
the end of this mail, except for APPS.AP_CANCEL_PKG, which remains invalid,
and running......
SQL select name, text from dba_errors where name='AP_CANCEL_PKG';
We get...........
NAME
------------------------------
TEXT
----------------------------------------------
AP_CANCEL_PKG
PLS-00201: identifier 'PA_INTEGRATION.PENDING_VI_ADJUSTMENTS_EXISTS' must be
declared
Does anyone out there have any ideas on the cause all these invalid objects,
and/or had experiences like ours???
We are running Applications 10.7, Oracle 7.3.3.5, Solaris 2.6
Any help would be much appreciated.
Jem.
The SQL to recompile invalid objects
Jem Lough
Switchboard 01865 330000
DDI 01865 333359
FAX 01865 791438
e-mail jem.lough@blackwell.co.uk
url http://www.blackwell.co.uk
Date: Tue, 20 Jul 1999 19:28:47 +0530
From: Gopalakrishnan K gopalak@wiproge.med.ge.com
Subject: RE: Invalid Objects
Hi Jem !
The invalid objects are caused by the structurual changes in the base
objects.( like a change in base table makes the view which is based on
the table invalid)
You have to run adadmin to recompile invalid objects to get rid
of invaild objects. sometimes you have to run twice/thrice since the
objcts have dependencies.
Regards !
Gopalakrishnan K
Database Administrator
GE Medical Systems Limited.
Pune 411014, INDIA
Date: Tue, 20 Jul 1999 21:03:51 PDT
From: "shankar govindan" shankargovindan@hotmail.com
Subject: RE: Invalid Objects
Gopal is right, that's why it happens. Most of the problems come up after
applying patches where in the base objects are modified and the related
proc,triggers,views and packages get disturbed and gets into the invalid
state. As we know it has to get recompiled when a user acesses it , but do
not know why that does'nt happen. Probably because of rights.
Anyway we apply patches normally after disconnecting the listener. This way
there are no locked objects. Then we run the script to check invalid
objects. We recompile at the backend using the same script.Then we use
adadmin and recompile once again. Check until it becomes Zero objects.Also
try recompiling Flexfield using adadmin if you are on the dev database.
I make it a standard practise to check these, believe it or not for 35 NT
APPS database and 8 HP-UX Database thru VSAT connectivity.
Shankar
APP-00001 Error in Oracle Inventory
Date: Thu, 22 Jul 1999 05:22:57 -0400 (EDT)
From: S F sf3010@yahoo.com
Subject: APP-00001 Error while defining workday calendar in Inventory
Hi all,
Our developers are getting the following error message in Oracle Apps.
10.7 Inventory module while defining Workday Calendar:
APP-00001 Cannot find message name MFG_SUBMIT_REQUEST
Cause: Message Dictionary cannot open the message file or cannot find
the message MFG_SUBMIT_REQUEST in the messge file.
Action: Check that your message file exists.
Any suggestion/clue to help is most welcome.
Best regards,
SF
sf3010@yahoo.com
Date: Mon, 26 Jul 1999 09:31:50 +1200
From: Erica Harris EricaH@wairc.govt.nz
Subject: RE: APP-00001 Error while defining workday calendar in Inventory
It sounds as though you may need to (re)generate the messages for MFG.
The program is FNDMDGEN which you can run from the command line. You can
also regenerate messages from adadmin if I remember rightly.
Let me know if you need more information and I'll go dig around in my
notes.
Regards,
Erica Harris
Date: Sun, 25 Jul 1999 20:53:52 PDT
From: "shankar govindan" shankargovindan@hotmail.com
To: oraapps-dba@cpa.qc.ca
Subject: RE: APP-00001 Error while defining workday calendar in Inventory
Harris is right we had similar problem once I remember and we regenerated
using adadmin.
Shankar
Trace Files
Date: Mon, 2 Aug 1999 09:48:23 -0500
From: "Bearden, Bill" Bill.Bearden@glenayre.com
Cc: "Tidd, Barry" Barry.Tidd@glenayre.com
Subject: Gen: Trace Files
I am posting this for a co-worker.
TIA,
Bill
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
We are getting an APP-18497 error from a form in OE. This occurred after
INV-C. Oracle support has asked us for a form level trace file and a
database trace file.
To get the database trace, I turned on trace from the App's Helptoolstrace
and produced a server side trace in /udump. No problem.
I then modifed the Winframe (client) sqlnet.ora to turn what I thought was
client side trace on. I sent this trace to Oracle and they rejected it.
First they said run it through TKPROF. They sent that output back and then
said the trace was in HEX format and they couldn't use it.
Anybody have any idea how to get a form level trace from the App to Oracle's
liking?
TIA from the Bible belt.
10.7.16.1 SC
HP UX 10.20
winframe for client level
ADUNLOAD utility 10.7 sc problems
Date: Sat, 7 Aug 1999 09:41:57 -0700
From: R Bhandarkar ak494@xoommail.com
Subject: ADUNLOAD utility 10.7 sc 16 problems
Our OS- AIX 4.3 Oracle 7.3.4, Apps 10.7 sc.
At present we have a/r fully installed. I want to convert
a/p from a shared to a complete install. I'm tring to use
adunload for this and having problems.
We have 2 machines apps and remote. Apps is where
oracle and evrything exists. Remote is the machine where the
cd is mounted so that apps can see it . I'm having problems
while using;
adunload product=ap from the $APPL_TOP directory.
I'm logged in the apps m/c as applmgr . do i need a unix id
applmgr with same password on the remote machin eas well ?
Is there are anything I have missed ?
Oracle support ha sbeen contacted , but their knowledge
level is depressing. I have also searched this group & found
nothing useful .
TIA
RAJA
Date: Sun, 8 Aug 1999 10:15:26 -0400
From: imbimbo_joe@adtranzna.com
Subject: Re: ADUNLOAD utility 10.7 sc 16 problems
What is the specific error you are getting. I am not an AIX expert, but in
SOLARIS, one first has to mount the CD on the remote machine and then
EXPORT it after the mount especially if some other CD was previously
mounted and exported. Without the error message, this is the obvious thing
to check.
After you export the CD filesystem on the remote machine, you have to NFS
mount it on the APPS server.
Joe
Adrelink error - Applying patch 708685 in inventory
Date: Mon, 16 Aug 1999 02:44:08 -0500
From: "Voss, Tracey Lee" tracey.voss@au.unisys.com
To: "'OraApps-L@cpa.qc.ca'" OraApps-L@cpa.qc.ca
Subject: adrelink error
Oracle Government Financials 10.7SC 16.7
RDBMS 7.3.4 / HP_UX 10.20
Hi all,
Having problems with an Oracle Applications patch (I'm sure we've all been
there before!) 708685 for Inventory
During adrelink stage get the following error:
Relinking module 'aiap' in product fnd ...
make -kf /d4s4/finroot/install/out/link_fnd.mk
/d4s4/finroot/fnd/6.1.1/bin/aiap
Starting link of fnd executable 'aiap' on Wed Aug 11 16:45:26 EST 1999
cc -s -t -L/d6s2/app/oracle/product/7.3.4/lib -o \
/d4s4/finroot/fnd/6.1.1/bin/aiap.new /d4s4/finroot/fnd/6.1.1/lib/fdamai.o \
/d4s4/finroot/fnd/6.1.1/lib/xitiap.o /d4s4/finroot/fnd/6.1.1/lib/xitusr.o \
/d4s4/finroot/fnd/6.1.1/lib/fdatma.o /d4s4/finroot/po/8.0.167/lib/libpo.a \
/d4s4/finroot/gl/9.0.9/lib/libgl.a /d4s4/finroot/alr/5.0.29/lib/libalr.a \
/d4s4/finroot/alr/5.0.29/lib/alstub.o /d4s4/finroot/rg/4.0.9/lib/librg.a \
/d4s4/finroot/pay/4.0.94/lib/libpay.a /d4s4/finroot/ff/4.0.94/lib/libff.a \
/d4s4/finroot/dt/3.0.94/lib/libdt.a /d4s4/finroot/ap/8.0.159/lib/libap.a \
/d4s4/finroot/as/1.3.38/lib/libas.a /d4s4/finroot/fa/7.0.166/lib/libfa.a \
/d4s4/finroot/pa/4.1.20/lib/libpa.a /d4s4/finroot/oe/4.0.172/lib/liboe.a \
/d4s4/finroot/ar/7.0.152/lib/libar.a /d4s4/finroot/bom/5.0.168/lib/libbom.a
\
/d4s4/finroot/eng/5.0.167/lib/libeng.a
/d4s4/finroot/crp/5.0.160/lib/libcrp.a \
/d4s4/finroot/inv/5.0.161/lib/libinv.a /d4s4/finroot/wip/5.1.1/lib/libwip.a
\
/d4s4/finroot/mrp/5.0.160/lib/libmrp.a
/d4s4/finroot/bom/5.0.168/lib/libbom.a \
/d4s4/finroot/mrp/5.0.160/lib/libmrp.a /d4s4/finroot/oe/4.0.172/lib/liboe.a
\
/d4s4/finroot/ax/1.2.23/lib/libax.a /d4s4/finroot/fnd/6.1.1/usrxit/libusr.a
\
/d4s4/finroot/fnd/6.1.1/lib/libfnd.a \
/d6s2/app/oracle/product/7.3.4/forms/lib/libforms.a -lsql -lsqlnet -lncr \
-lsqlnet -lclient -lcommon -lgeneric -lsqlnet -lncr -lsqlnet -lclient
-lcommon \
-lgeneric /d6s2/app/oracle/product/7.3.4/lib/libepc.a -lnlsrtl3 -lc3v6
-lcore3 \
-lnlsrtl3 -lcore3 -lnlsrtl3 `cat \
/d6s2/app/oracle/product/7.3.4/rdbms/lib/sysliblist` -lm
Undefined first referenced
symbol in file
bmfrlst_free_component_list /d4s4/finroot/wip/5.1.1/lib/libwip.a(wilmeb.o)
bmaedi_ato_expl_dem_interface
/d4s4/finroot/inv/5.0.161/lib/libinv.a(inldbm.o)
bmslxpl_explode_single_level /d4s4/finroot/wip/5.1.1/lib/libwip.a(wilmeb.o)
bmgetrvn_get_revision /d4s4/finroot/po/8.0.167/lib/libpo.a(rvexr.o)
cmlwdb /d4s4/finroot/wip/5.1.1/lib/libwip.a(wiluur.o)
bmlcrdm_cr_config_demand_row /d4s4/finroot/oe/4.0.172/lib/liboe.a(oeamrl.o)
bmaed_ato_expl_demand /d4s4/finroot/inv/5.0.161/lib/libinv.a(inlddm.o)
bmldlci_delink_config_item /d4s4/finroot/oe/4.0.172/lib/liboe.a(oeagrp.o)
UX:ld: ERROR: /d4s4/finroot/fnd/6.1.1/bin/aiap.new: fatal error: Symbol \
referencing errors. No output written to
/d4s4/finroot/fnd/6.1.1/bin/aiap.new
*** Error code 1 (bu21)
Done with link of fnd executable 'aiap' on Wed Aug 11 16:45:28 EST 1999
Done relinking module aiap in product fnd
This happens for a number of other executables as well. Previous to this
patch I had applied three patches without incident.
Any help is much appreciated!
Tracey Voss
Applications Support
Unisys Australia
ph: (02) 6274 3530
fax: (02) 6274 3533
Date: Mon, 16 Aug 1999 09:12:03 +0100
From: Graham Duggan gjduggan@mail.com
To: "'oraapps-l@cpa.qc.ca'" oraapps-l@cpa.qc.ca
Subject: RE: adrelink error
I am not on HPUX however the code structure is similar enough across
platforms to make some analysis. I searched for some of these missing
libraries on my Solaris box and found (for the four I searched for) that
they exist in
/prod/dapplmgr/r10.7/bom/5.0.168/lib/libbom.a
[35] | 10696| 60|FUNC |GLOB |0 |2
|bmfrlst_free_component_list
[18] | 16| 1112|FUNC |GLOB |0 |2
|bmaedi_ato_expl_dem_inter
[41] | 9632| 7680|FUNC |GLOB |0 |2
|bmldlci_delink_config_item
[21] | 16| 10664|FUNC |GLOB |0 |2
|bmslxpl_explode_single_level
I see that BOM_TOP/lib/libbom.a is in the libraries included in the
relink. I suggest that there must, since this latest patch, be a problem
with this library on your system.
Did the first part of this patch apply successfully on your system?
Was this library modified during the application of this patch?
Did you run out of space while this library was being modified?
Are you familiar with the ar and nm command? read the man entries for
these command to see if you can rebuild the symbol table for the
library.
Call support. There may be an open issue on the patch you are applying.
Hope this helps
Graham.
Date: Mon, 16 Aug 1999 07:56:08 -0700
From: Karen Blackwell KBlackwell@rockshox.com
Subject: RE: adrelink error
During my first pass at this patch, I had numerous problems also. My notes
indicate:
At errors, fix invalid objects then continue
Validate all invalid objects when complete
Ensure all listed executables (in readme.txt) are relinked (check
date/time stamp and re-link manually as needed)
(I am using Sun Solaris on 10.7NCA and 7.3.3.6)
HTH
Adrelink errors on application of patch 717183
Date: Mon, 16 Aug 1999 13:05:09 -0400
From: "Avrami, Louis" louis_avrami@es.adp.com
To: "'oraapps-dba@cpa.qc.ca'" oraapps-dba@cpa.qc.ca,
Subject: GL: Error when running adrelink
Hello all,
I had a problem generating GL executables when trying to run a
adrelink force=y ranlib=y all
when applying patch 717183 on our Release 10.7 Smartclient 16.1 AIX
4.3.2 installation.
The patch recommended doing a adrelink on everything after it was
installed. When I ran adrelink, all the executables for all of the
fully-installed products except for GL were generated. Below is the
failure in the $APPL_TOP/install/log/adrelink.log file.
What is interesting is that an 'mcs' and then a 'strip' is attempted
for each GL executable. For all other products, only an 'mcs' is done. The
makefile fails with an error code of 255 on the first strip, so no GL
executables are generated. What is interesting is that previous adrelinks
did not fail on the strip, they would ignore it:
mcs /d04/app/applmgr/10.7/gl/9.0.9/bin/GLXFMA
strip /d04/app/applmgr/10.7/gl/9.0.9/bin/GLXFMA
strip: /d04/app/applmgr/10.7/gl/9.0.9/bin/GLXFMA -- 0654-420 The file was \
already stripped as specified.
make: The error code from the last command is 255.
(continuing)
Done with link of gl executable 'GLXFMA' on Sat Aug 14 06:29:46 EDT 1999
Has anyone else experienced a similar problem? Oracle Support
hasn't been very helpful so far. It appears to me that there is some
problem with the GL makefile that needs to be corrected.
Lou Avrami ( louis_avrami@es.adp.com )
10.7NCA: NT-client Apps not loading
Date: Thu, 19 Aug 1999 13:24:46 -0400
From: Kathy Largay klargay@bhe.com
To: "'oraapps-l@cpa.qc.ca'" oraapps-l@cpa.qc.ca
Subject: 10.7NCA: NT-client
We are currently running 10.7 NCA on NT clients with the Oracle Applications
Server. We have production and test instances which run fine (we use
Jinitiator). We now have a development instance, and we configured a third
NT server to access it. When I run Jinitiator, it loads the applet, but
then does nothing else -- it doesn't bring up the Oracle sign-on screen.
Has anyone run in to this before? I've called Oracle Support, and didn't
get much of an answer. Thanks!
Oracle 8.04 and Smartclient together gives GPF
Date: Thu, 19 Aug 1999 14:13:29 -0400
From: mnatale@manu.com
To: oraapps-l@cpa.qc.ca
Subject: SmartClient - fnd75win.dll error - URGENT!
In the past, we've received the generic GPF error in fnd75win.dll and been
able
to resolve it by applying appropriate client patches and adding resources
(memory, disk space, etc.). However, this is a new problem:
Our company needs to support our own products on Oracle v 8.0.4 on
Windows NT
4.0, Service Pack 5.
We've installed Oracle 8.0.4 into ORANT 32-bit Oracle home.
We've installed 10.7 SmartClient into ORAWIN 16-bit Oracle home as usual.
After we loaded the SmartClient piece, we receive the following error
immediately after launching the Oracle10SC Smart Client Applications. We
don't
even get to the login screen.
F45RUN caused a General Protection fault in module FND75WIN.DLL at
0001:0B79.
We have open a TAR with Oracle , but wonder if anyone has been able to get
Oracle8.0.4 and Smartclient to co-exist?
Thanks!
Maria Natale Beem
Manugistics, Inc.
2115 East Jefferson Street
Rockville, MD 20852
Oracle Payroll process is very slow
Date: Thu, 22 Jul 1999 10:29:38 -0500
Date: Thu, 22 Jul 1999 16:06:26 -0400 Distributed Environment - Apps in
Date: Tue, 27 Jul 1999 11:36:57 +0500 Performance issues - V. 11.02
From: Greg Grimes[SMTP:gpgrimes@gte.net]
From: Margaret Murray [mailto:mmurray@delorme.com]
Date: Fri, 14 May 1999 07:14:55 -0500
Date: Tue, 15 Jun 1999 11:13:17 -0400
Date: Tue, 15 Jun 1999 12:03:47 -0400
Date: Tue, 15 Jun 1999 17:15:47 -0400
Date: Mon, 21 Jun 1999 17:40:11 -0400
Date: Tue, 22 Jun 1999 11:11:22 -0400
Date: Tue, 22 Jun 1999 11:31:45 -0400
Date: Tue, 22 Jun 1999 12:59:23 -0400
To: Darlene D Dougherty/HQ/CP/CNG@CNGNOTES
Date: Tue, 22 Jun 1999 15:54:44 -0400
Date: Tue, 22 Jun 1999 15:30:05 -0500
From: Kurt Greener kgreener@styline.com
To: "Oracle DBA list (E-mail)"
Subject: Oracle Payroll
We are currently running Oracle Payroll for approx. 750
employees. The payroll process takes from 1.5 - 2 hours. Does anyone
else run Oracle payroll? Does it take that long for you?
We run HP-UX 10.20 on a dual processor K9000 as the dB server.
The database is 7.3.4.0. The applications are 10.7NCA. The app server
is NT.
While payroll was running sar -d showed;
Please click here for details
From: gfury@manu.com
To: oraapps-dba@cpa.qc.ca
Subject: Re: Oracle Payroll
Looks to me like c0t6d0 is quite busy. What is on it? Data files? redo
logs?
If they are data files make sure that you have separate tablespaces for
data and indexes
and you place them on different drives.
Do you only have 3 disks on this system? in order to run an Oracle
database with good
throughput you need at least 7 disks.
Look at the hit ratio of the database, you may need to increase the size of
db_block_buffers.
Greg Fury
OCP DBA
Manugistics, Inc
Rockville, MD
From: Irfan Haroon Seikha irfan@jaffer.com
To: Multiple recipients of list
Subject: Oracle Applications in Distributed Environment
Hi All,
I like to know if this is possible in Oracle Apps.
Location A,C and D Apps installation using PO,Inventory, General Ledger
(Database Servers A/C/D)
Location B Apps installation General Ledger,Inventory,Fixed Assests.
(Database server B).
Each location is in different City. Location B is the Headquarter and
location A is Regional Office. Th
If yes then what type of expertise are required in term of Database
Administration.
Regards
Irfan
Irfan Haroon Seikha
Jaffer Consulting
Sent: Thursday, May 13, 1999 10:49 AM
Subject: Performance Issues with 11.02
I would a appreciate any performance tips concerning Release 11.02.
Database, Server, Middle tier, whatever would be helpful.
Thanks.
Sent: Thursday, May 13, 1999 3:07 PM
Subject: RE: Performance Issues with 11.02
What is your hardware? We have Solaris for both middle and db server; so of
these suggestions might not make sense on NT.
Large swap space on middle and db tier, minimum 3xRAM, higher if possible.
Lots of CPU and RAM (as much as you can afford) on the client; Oracle
minimums are not adequate (only thin part of client is the hard drive).
Largest network bandwidth you can afford between tiers (especially between
middle and db server).
Margaret Murray, DBA
DeLorme Mapping
From: Bharat Patel bpatel@DOMINOAMJET.com
Subject: RE: Performance Issues with 11.02
We have middle and db server on HP-UX 11.0 with two CPUs and 2GB RAM and 8GB
swap space which is 2GB more then oracle recommandation.We don't have yet
any performance problem.Also to get a good performance you need to get rid
of the *.DAT file in OTRACE directory on DB server as it is potential
performance problem. Due to this it really slow down the system.
Bharat Patel
D.B.A.
bpatel@dominoamjet.com
847-244-2501 ex 1249
From: NSmith@micros.com
Subject: Re[2]: Performance Issues with 11.02
Bharat,
I'm curious about your instance. We have upgraged from Oracle Apps
10.7 to 11.02 and are having serious performance problems.
We have the following:
Oracle Apps 11.02 (AP,AR,GL,PO,MRP,BOM,OE,OSM,CN,INV, etc)
Oracle Server 8.1.5.1
HP-UX 11.0
HP 9000/K570
6 CPUs
5 GB RAM
6 GB swap space
Approximately 200 to 300 users.
How do you compare with our modules and users?
HP experts say we are correctly sized and configured...they don't see
why the Oracle Apps aren't "screaming". Any ideas would be
appreciated.
Thanks,
Neil Smith
From: "Mark W. Farnham" mwf@rsiz.com
Subject: RE: Re[2]: Performance Issues with 11.02
You've made no mention of your disk farm. Having sufficient capacity for
i/o's per second and aggreggate disk farm throughput is necessary but not
enough, especially for Oracle Applications. Careful attention must be paid
to object placement tuning for large Oracle Applications installations to
perform well.
You have also only mentioned the database server. Is that where your
problem lies? If you place a serious hot rod PC in place of one that is
marginal, do things get better? If you don't have spare machines to move
around, then see if there is a difference if you just disable screen savers
and wallpaper. If either of these tests improves performance, then the DB
is not your worst problem.
If you place the same PC on various network connections (including a
private segment in the the same room with the applications and db servers),
does the performance improve as you get closer to the machine? Then
overloaded segments, hubs, routers, switches, or latency could be the
problem. If you have a satellite link any way along the line: Give up until
that link is gone.
Is the problem only a slow signon and then everything is okay?
Is the problem only moving from form to form but okay within a single form?
Is the problem with large batch jobs such as open gl period?
To productively make progress on performance improvement it is essential to
identify the worst bottleneck for a particular problem. Then making that
bottleneck of sufficient throughput will create the possibility of finding
the next limitation.
Comparing the basic DB server machines still may be illustrative if you are
grossly undersized. Sixty users per CPU does seem a little high to me,
depending on the chip rate and batch load. However, based on different load
requirements and reconfiguring something such as the number of code
combinations in GL I have seen over an order of magnitude difference in
throughput for the same hardware and user community. That is why careful
isolation and correction of worst bottleneck first is important.
Good luck,
Mark
From: "Amin Adatia" amin@knowtech.on.ca
Subject: Re[2]: Performance Issues with 11.02
Neil,
If you have RAM of 5GB then the SWAP should be greater than 10GB. Better
would be 3x RAM. How about the block size, db_block_buffers, etc
parameters?
Regards
Amin Adatia (KnowTech Solutions Inc www.knowtech.on.ca )
amin@knowtech.on.ca
(613) 226 8378 Fax (613) 226 7004 Cellular (613) 282 8300
From: NSmith@micros.com
Subject: Re[3]: SOLUTION: Performance Issues with 11.02
All,
Thanks for all the responses. Oracle solved the problem.
TWO SOLUTIONS:
1. VERY IMPORTANT: The user interface (UI) will not perform
optimally with display settings higher than "256 color" and "800x600
dimensions". Not having these settings can cause serious UI
performance issues (i.e. delays in displaying screens and moving
between fields). I suppose that pumping up the video RAM may be an
alternative solution. Users are happier now.
2. If you have implemented the Self-Serve Apps, you may encounter
some huge runaway processes. They are intermitant and can be huge
resource hogs. Ours were taking upwards to 0.5 G of RAM and sucking
down all available CPU. Oracle is still working on a patch...we're
still getting them but at least we know what to look for and kill
them. The following SQL is their fingerprint (along with resource
hogging):
begin if :first=1 then owa.init_cgi_env(:numCGIParams, :cgiArgn,
:cgiArgv); htp.HTBUF_LEN := 256; oracleapps.vl(I_1=:1,I_2=:2);end
if; owa.get_page(:lines,:nlines);end;
Hope this helps.
Neil
From: "Michael Porter" MPorter@pgahq.com
Subject: Re: Re[3]: SOLUTION: Performance Issues with 11.02
Is this potentially true for 10.7/16.1 smart client performance as well???
The user interface (UI) will not perform
optimally with display settings higher than "256 color" and "800x600
dimensions". Not having these settings can cause serious UI
performance issues (i.e. delays in display
TIA,
Mike
From: NSmith@micros.com
Subject: Re[5]: SOLUTION: Performance Issues with 11.02
Mike,
We upgraded from 10.7 to 11.02. The users had been used to SC and the
new UI for 11.02 was significantly slower. I don't know if this could
help SC but give it a try on a few PC's and see.
Neil
From: Darlene_D_Dougherty@CORP.CNG.COM
Subject: Re: Re[3]: SOLUTION: Performance Issues with 11.02
the runaway processes can be fixed via patch # 648643,,, the runaway processes
are caused via users not exiting via the apps and the cookie does not
reestablish itself,, the patch will ensure a new cookie every time...it has
improved our performance tremendously
cc: Chris Blais CBlais@ixc-comm.com
Subject: Patch 648643 is incorrect as a performance fix
I believe your patch 648643 refers to another (older) problem and not the
one listed.....
Do u have any other info?!?
The patch # is NOT a fix, but a bulletin...... (there is no fix for having
multiple Netscape sessions ( and causing performance problems) b/c Oracle
officially does not support 2 Netscape sessions, (hence the fix)
Regards,
Chris Blais
DBA consultant
From: Darlene_D_Dougherty@CORP.CNG.COM
To: Chris Blais CBlais@ixc-comm.com
Subject: Re: Patch 648643 is incorrect as a performance fix
yes the patch # I gave is correct for 10.7 prod 16.1 sequent dynix , the tar #
they downloaded the patch to is 11994574.600
the patch resolves the issue of a new cookie not being spawned if there is an
existing cookie(hence the existing limits are retained ,, the patch forces a
new cookie upon each session you open.. what the patch does is limit you to
having only one session open at a time.
From: Chris Blais CBlais@ixc-comm.com
To: "'Darlene_D_Dougherty@CORP.CNG.COM'"
Subject: RE: Patch 648643 is incorrect as a performance fix
To clarify your response to "performance fix for ICX (self-service apps)"
1. Patch 648643 is for 10.7 only. The problem was fixed in Rel. 11.0.x.
2. Oracle does not support multiple Netscape logins from the same PC
3. Your correct, the patch is the right number for the 10.7 problem.
Thanks for your update,
Chris Blais