3.0.9.8.3 Issues on Tru64

Hey,
I have just about managed to get the 30984 upgrade complete on a Tru64 instance of portal. However I do still have one invalid portal30 package. This package is causing any forms in portal to be unusable.
The Package is wwa_api_module_event, when I attempt to recompile it I get rhe following error:
LINE/COL ERROR
0/0 PLS-00707: unsupported construct or internal error [2603]
This would finish off the upgrade and migration for us.
Does anyone have any ideas about why this is happening ? I am currently talking with support but have not been able to come up with a solution.
thanks

===========================================
Oracle Portal Patch Release 3.0.9.8.2 For Solaris Platform
===========================================
Portal Patch Release 3.0.9.8.2 for Solaris is now available for
download from metalink. Bug Number for the Patch Release is:
2104468.
This can be applied to Release 3.0.9 of Oracle Portal
(Oracle9iAS v1.0.2.2). You can also apply this over a Portal
with 3.0.9.8.1. This Patch Release is cumulative. When applied
to 3.0.9.8.0 it applies the earlier Patch Release (3.0.9.8.1) as
well.
The installation instructions are in README.TXT and RELNOTES.HTM
contained in the Patch Release.
To access additional information about the patch (including bugs
fixed), review the release notes accessible on Portal Studio at
http://portalstudio.oracle.com (click the link for Support
Center).
NT release and other ports will be available soon (week or so).

Similar Messages

Installation problem of Weblogic 5.1 on Tru64 5.1

It has been mentioned in Release notes that Weblogic 5.1 works with jdk1.2.2
on Tru64 Unix version 5.1.
The following were the activities carried out on Compaq Tru64 OS :
1.Weblogic 5.1 server and Weblogic 5.1 SP8 has been downloaded (specifying the
option for non Win32 users while downloading) from BEA website and unzipped
using the jar utility.
2.OS version used is
$ uname -a
OSF1 xp1000 V5.1 732 alpha
3.java version used is "1.2.2-6"
Classic VM (build J2SDK.v.1.2.2:04/26/2000-09:58, native threads, jit_122).
4.The PATH and LD_LIBRARY_PATH environmental variables are set to :
echo $PATH
.:/usr/bin:/usr/opt/java122/bin
$ echo $LD_LIBRARY_PATH
/usr/lib:/weblogic/lib/tru64unix:/weblogic/lib:/usr/opt/java122/bin
5.Java class path has been updated to include Path of weblogic510sp8boot.jar
and Weblogic class path has been updated to include weblogic510sp8.jar as
recommended in the service pack release notes.
6.The options recommended (to resolve known issues with Compaq Alpha on
Weblogic 5.1)in BEA site has also been taken care in java startup
script in startWebLogic.sh.
Known Issues with Compaq Alpha on WebLogic 5.1 (as in Bea site)
JDK 1.2.2 is a supported JVM when using Weblogic 5.1 on a Tru64 Unix platform.
However, the WebLogic Server's startup script needs to be modified in order
for the server to start properly. The JIT(Just-In-Time compiler) must be
turned off; whereby the -Djava.compiler directive needs to be passed to the
JVM and the JIT disabled.
For example:
/usr/opt/java122/bin/java -Djava.compiler= -ms128m -mx128m -classpath
./classes/boot -Dweblogic.class.path=./license:./classes:./lib/weblogicaux.jar:
./myserver/serverclasses -Djava.security.manager -Djava.security.policy==
weblogic.policy weblogic.Server
The error observed while starting the script startweblogic.sh is :
$ ./startWebLogic.sh
LD_LIST=/weblogic/lib/tru64unix
2224:/bin/ksh: /sbin/loader: Fatal Error: Cannot map library /weblogic/lib/tru64unix
7. The files existent in this directory are :
weblogic/lib/tru64unix
$ ls
libmuxer.so libwlenv.so zac_tru64unix
libproxy.so mod_wl.so zac_tru64unix_g
libweblogicunix1.so oci805_7
[problem.txt]

Can you post your startWebLogic.sh script?
BALAKRISHNAN GANESH wrote:
It has been mentioned in Release notes that Weblogic 5.1 works with jdk1.2.2
on Tru64 Unix version 5.1.
The following were the activities carried out on Compaq Tru64 OS :
1.Weblogic 5.1 server and Weblogic 5.1 SP8 has been downloaded (specifying the
option for non Win32 users while downloading) from BEA website and unzipped
using the jar utility.
2.OS version used is
$ uname -a
OSF1 xp1000 V5.1 732 alpha
3.java version used is "1.2.2-6"
Classic VM (build J2SDK.v.1.2.2:04/26/2000-09:58, native threads, jit_122).
4.The PATH and LD_LIBRARY_PATH environmental variables are set to :
echo $PATH
.:/usr/bin:/usr/opt/java122/bin
$ echo $LD_LIBRARY_PATH
/usr/lib:/weblogic/lib/tru64unix:/weblogic/lib:/usr/opt/java122/bin
5.Java class path has been updated to include Path of weblogic510sp8boot.jar
and Weblogic class path has been updated to include weblogic510sp8.jar as
recommended in the service pack release notes.
6.The options recommended (to resolve known issues with Compaq Alpha on
Weblogic 5.1)in BEA site has also been taken care in java startup
script in startWebLogic.sh.
Known Issues with Compaq Alpha on WebLogic 5.1 (as in Bea site)
JDK 1.2.2 is a supported JVM when using Weblogic 5.1 on a Tru64 Unix platform.
However, the WebLogic Server's startup script needs to be modified in order
for the server to start properly. The JIT(Just-In-Time compiler) must be
turned off; whereby the -Djava.compiler directive needs to be passed to the
JVM and the JIT disabled.
For example:
/usr/opt/java122/bin/java -Djava.compiler= -ms128m -mx128m -classpath
/classes/boot -Dweblogic.class.path=./license:./classes:./lib/weblogicaux.jar:
/myserver/serverclasses -Djava.security.manager -Djava.security.policy==
weblogic.policy weblogic.Server
The error observed while starting the script startweblogic.sh is :
$ ./startWebLogic.sh
LD_LIST=/weblogic/lib/tru64unix
2224:/bin/ksh: /sbin/loader: Fatal Error: Cannot map library /weblogic/lib/tru64unix
7. The files existent in this directory are :
weblogic/lib/tru64unix
$ ls
libmuxer.so libwlenv.so zac_tru64unix
libproxy.so mod_wl.so zac_tru64unix_g
libweblogicunix1.so oci805_7
Name: problem.txt
problem.txt Type: Plain Text (text/plain)
Encoding: base64

Invoices not printing since migration of AppsTier from HP Tru64 to Linux

We have migrated our Appps Tier from HP Tru64 unix to Linux server and since the migration our concurrent requests "Invoice Print Selected Invoices" are getting "completed warning", when I look at the log it says the following error:
Printing output file.
Request ID : 940190
Number of copies : 1
Printer : lp4
/bin/sh: /usr/local/bin/enscript: No such file or directory
APP-FND-00500: AFPPRN received a return code of failure from routine FDUPRN. Pro
gram exited with status 127
Cause: AFPPRN received a return code of failure from the OSD routine FDUPRN. Pro
gram exited with status 127.
Action: Review your concurrent request log file for more detailed information.
I wonder why it is trying to open a non existant directory /usr/local/bin/enscript ?
Can someone help.
Regards
Prashanthi

Try to relink adadmin as follows, and check then:
- Source the application env file
- Issue the following:
$ adrelink.sh force=y "ad adadmin"

Getting error during the Oracle9i installation "Oracle9i Database Release 2 (9.2.0.1.0) for Compaq Tru64 machine"

Hi All,
Getting follwing error during the Oracle9i installation “Oracle9i Database Release 2 (9.2.0.1.0) for Compaq Tru64 machine"
u01/app/oracle/product/9.2.0/install/make.log
- Linking Oracle
rm -f /u01/app/oracle/product/9.2.0/rdbms/lib/oracle
cc -L/u01/app/oracle/product/9.2.0/lib/ -L/u01/app/oracle/product/9.2.0/rdbms/lib -o /u01/app/oracle/product/9.2.0/rdbms/lib/oracla
ld:
load_whole_file: Could not malloc 31328. bytes, for /u01/app/oracle/product/9.2.0/lib//libn9.a(asndata.o), didnt attempt to mmap
Stop.
*** Exit 1
While checked the memory details in the machine its showing plenty of free virtual memory.
# ps aux
USER PID %CPU %MEM   VSZ RSS TTY      S STARTED         TIME COMMAND
oracle     1353 0.6 27.2 601M 555M pts/2    S N 09:14:07 5:49.67 /tmp/OraInstall2014-01-21_09-14-02AM/jre/bin/../bin/alpha/native_threads/java -Doracle.installe
root       3033 0.0 0.4 10.4M 7.7M pts/0    R + 10:59:44 0:00.06 ps aux
oracle     1351 0.0 0.0 2.46M 320K pts/2    I N 09:14:02 0:01.36 ksh /tmp/OraInstall2014-01-21_09-14-02AM/installer -Doracle.installer.library_loc=/tmp/OraInsta
oracle     1277 0.0 0.0 2.48M 312K pts/2    I + 09:02:44 0:00.04 -ksh (ksh)
root       1269 0.0 0.0 2.56M 232K pts/2    I    09:01:23 0:00.01 -sh (sh)
root       1268 0.0 0.0 5.31M 768K ??       I    09:01:22 0:00.11 /usr/bin/X11/xterm -ls
root 1 0.0 0.0 608K 96K ?? IL   08:45:31     0:00.08 /sbin/init -a
root 0 0.0 3.9 2.52G 79M ?? R < 08:45:31     0:23.16 [kernel idle]
And I have rebooted the machine still also getting the same error.
Is there any way to resolve this issue??
Regards
prasanta

Pl elaborate the business / technical reason for installing a 15+ year old unsupported software on a 15+ year old unsupported hardware, assuming you are installing an a supported OS version
HTH
Srini

9iAS Installation on HP Tru64 Cluster

We have sucessfully managed to install both the Infrastructure and middle-tier on a Tru 64 Cluster. The application (forms 9i) runs perfectly OK, however, we cannot get the report server to work. (report server called reportserver rather than rep_apollo)
We have a development environment, which is also running Tru64 which runs perfectly OK.
We have the osagent installed on an NT machine on the same sub-net, and works OK for development.
On the Cluster machine, if we go to the middle-tier's $ORACLE_HOME/vbroker4/bin directory and type osfind, nothing is found. If we set the environment variable OSAGENT_ADDR to the machine running the osagent, and then perform an osfind it works OK - we have therefore set the environment variable in the profile, reports.sh and rwengine.sh but still the reportserver does not register with the osagent
Is there any other file that we need to change for the reportserver engine to know where the osagent is ?
Thanks in advance

Richard,
We have also installed 9iAS on a non-clustered Tru64 box. Sometimes the reports server starts, mostly not the first time, stays up for no more than 24 hours and dies. We are using version 9.0.2.3 (fully patched) on Tru64 5.1a box.
I have logged numerous TARs but not getting andy real answers. The osfind problem you reported i believe is spurious because the osgaent we have running on a windows box as the Visibroker people never wrote it for Tru64, don't know if this is still the case for 9.0.4 which is only certified on 5.1b. If i issue the osfind on windows it's running!
We are going to try 9.0.4 on a windows 2000 box and if it works then we will go done that path. Personally i feel the problem is with 9iAS not communicating with the osagent on windows properly?
Regards,
Trent

Database migration from Tru64 to HPUX on 10g

We'll be migrating off of Tru64 to HP-UX on Oracle 10g. We'll be using data pump to export the Tru64 version and data pump to import into the HP-UX release. All of the documentation that we've found says to do a full export and a full import. Full exports include the SYSTEM schema in addition to the application owner. Our plans are to do the following:
1) perform a full export from the Tru64 server
2) create a new instance on the HP-UX side, that includes creating all of the tablespaces
3) perform a full import into the HP-UX side
Because we have performed a full export, when we do a full import we get warnings about the SYSTEM objects existing. What do we do about the warnings?
What is the best practice for full database imports using data pump?
Are there other considerations we should be aware of?
Thanks in advance.

I remember system schema issues from the old import and export, there is normally no problem from memory .. but i am thinking from memory.
Perhaps worth noting the rowcounts and objects before and after your import just in case anything new has appeared ... i seem to remember some related to sqlplus.
I always think exports are easy ... imports are harder.
Also if your imports are giving you time problems there may be some things that can help speed it up: There are others who could give better advice on this.
See for example here [http://www.dba-oracle.com/t_import_impdp_import_performance_speed.htm] ... i came accross this with a quick google.
Tuning a temporarily large db_cache_size / sga_target / pga_target might be helpful, but if paging is caused this is counter productive.
The parallel option on impdp might be helpful.
The following are complicated and probably should only be tried if you need the speed:
- Temporarily setting larger tables with more than one index to cache before export and changing back afterwards.
- To go really on the knife-edge move your redo logs to a ramdisk temporarily to get a little extra speed , but this is a very very very dangerous practice. (I am nutter who has done this!)

Production environment in Linux, development testing in Tru64

We are going to be upgrading to 9i from 8i and will need to purchase new hardware. So our possible technical architecture in the future will look like this: Production DB and 9iAS running on separate Linux servers, while Development and testing still located on Tru64 (DB) and WinNT (AS). Does anyone know of any compatibility issues with developing in a Tru64 environment and releasing onto a Linux environment.

Hi,
It all depends on your infrastructure and what products you are using.
In most cases dev/prod are on different servers and are separate installs, this is the route I would promote.
Some clients do share an environment for dev/test but not all, not everybody is the same.
Cheers
John
http://john-goodwin.blogspot.com/

Abnormal memory allocation in RMI Server on Unix tru64

Hi all,
I have a problem with an RMI server application that run on Unix tru64 v. 5.1b (jvm 1.4.2).
I start the application in the following way:
1) start rmiregistry with the options -J-Drmi.server.hostname=myhostname myPort.
2) Start my class StarterRmiProcesses which have a main method in which create four new instance of different RMI objects (the instance are static class attributes) and registry those by:
Naming.rebind("rmi://" + hostName + ":" + portNumber + "/RmiObjectName", rmiObject).
The problem is that at the startup jvm allocate about 98 MB of memory, and when a client connect to the rmi server increase the memory usage by 1 MB, I have monitored the application for 2 or 3 days and sow that the memory usage always
increase and never decrease.
I tried to run the same application on Windows XP (in the same way if except that I don't start before the rmiregistry but use the LocateRegistry.createRegistry(myPort) ), and I sow that the initial allocation is about 20 MB, when a client connect to the rmi server increase the memory usage by 1 MB too but periodically the memosy usage decrease(I suppose when the gc run).
Any suggestion?????
Thanks in advance.

At a guess this is a JVM issue not an RMI issue, as the RMI code being executed is almost entirely the same on both platforms.

Integrating Bank Wizard and Forms 6 on Tru64?

Hi all,
I'm currently working at a site that is thinking about trying to call Bank Wizard from Eiger Systems from an old character based Forms 6 application that is running on Tru64.
Does anyone know if this is even possible?
Better still, if you have any experience of this sort of integration, what were the problems that you faced?
All thoughts would be greatly appreciated,
a.

Hello;
This is an intersting issue;
Could you please gave me an adress
for web site talking and including same simples
for OLE.
Regards

Portability issues

Hi,
We could not build db-4.5.20 "out of the box" on solaris2.6 or 2.7 or on AIX-4.3, hpux10.20, or Tru64 (4,0D and 5.1) with the native compilers. Most of the problems were to do with socklen_t, but there were one or two other little things. We have patches, but I am not sure where to send them.
Peter

Hi,
I also had problems building db-4.5.20 on Tru64 5.1. A few of the repmgr modules in particular would not compile because of socklen_t issues. Are the patches referred to previously available for download?
Kathy

Moving from 10gR2 Db on Unix (Tru64) to a new Linux server (10gR2)

Hi,
We are planning on moving our oracle 10gR2 Db on Unix (Tru64) to a new Linux server (same database 10gR2). Interested in discussing the easiest way for a newbie to get the data over to the new box. We currently perform a 'COLD" backup and have no transaction logging. The database is a data warehouse so stopping the database and backing up the files is not a issue. The database size is 500GB.
The Oracle ADDM on our current box is showing "individual database segments responsible for significant user I/O". There is a tablespace with about 30 separate data files that need to be "reorganized"!
Should we perform the reorganize before the migration or after on the new faster box?
What is the easiest way to do this for a newbie? Thanks.

Since your source and target platform all little ENDIAN format. Cold copy from OS level is fastest and easiest way.
If you plan to reorg the tablespace along the way, you have to do export/import, use data pump since you are on 10g.

Installing Oracle 734 on tru64

Hi All
After a h/w issue ("the disk-controller blew taking out the os but maybe not the array" - whatever that means :-) ) , the o/s is being rebuilt, and I may soon be called upon to re-install Oracle 734 on tru64 (circa 1999).
My knowledge is Oracle9i on Solaris (argh!!)
please help.
Richard101

I am not pretending to write a full reference, the documents listed above are enough for a 7.3 administration. Oracle 9i/10g has a graphical environment. Back in 7.3 everything was CLI (command line interface). You have to manually create everything, database, listener.ora, tnsnames.ora, sqlnet.ora files.
It didn't have sqlplus as the administrative tool, it was jus a client you can use to connect regular users. The connect / as sysdba connection didn't exist, there in only was the svrmgrl CLI to perform administrative tasks, such as recover, create database, and all regular SYSDBA tasks.
There was no spfile, but the regular init.ora file. Just to name some meaningful first hand features.

Can't find class error on Tru64 UNIX

I have written a class, TextProcessor, and I can compile and run it without any problems on my Win2K box. I ftp the source to works Tru64 box and compile the class just fine. When I issue the command "java TextProcessor" I get the msg "Can't find class TextProcessor". I immediately thought I had a classpath issue. I messed around with setting my classpath but without luck.
I then wrote a new class on the Unix box using vi called Hello.java. I compile the Hello class without a problem. I issue the command "java Hello" and get "Hello!", just like I am supposed to get.
The 2 classes reside in the same directory, TextProcessor does not have any package statements. I am totaly confused.
Environment information:
Tru64 UNIX version 5 or so.
JDK 1.1.8 (not my choice and I have no control over the JDK)
JRE 1.2.2 is also installed.
Please help!
Thanks

Maybe.....
Make sure that you have the permissions necesary to read,write and execute the file from your own account.
If you canno't find any class maybe is because you cannot read.
What account are you used to pute the file from Win?
you can try this...
chmod 777 TextProcessor.java
chmod 777 TextProcessor.class
But......????? uhmmmmmm??
I'm not sure, because If you can compile then you can read the file, but it's depends you umask ......
it's only a suggest

BC4J app deploy on Tru64 HW

Hi all,
as in subject my BC4J/JSP application (made with JDev 9.0.2.727) should be deployed on a Compaq Tru64 4.0 machine. Are there issues and known problems about it?
TIA,
Lucio Pitocco

No BC4J-specific issues that we know about in this configuration.
Make sure to upgrade to the production 9.0.2.829 release. The release you mention is not production.

Problem with libclntsh.so on Tru64

Hello,
We are developping a C application using OCI 7.3 Data access fonctions on Unix servers.
One of our clients is using an Alphaserver with Tru64 5.1 and an Oracle 8.1.7 database. Our application used to work correctly on this server. But recently, this client installed a patch kit (t64v51b19as0005-20020411) from Compaq (who asked for it for some supports reasons), and since that our application doesn't work at all.
When we launch the application, the following message immediatly occurs:
Illegal instruction(coredump)
After some deeper tests, it appeared the problem is coming from the library libclntsh.so which seems to be non-compatible with the new kernel (the problem occurs only when the library is linked with the application). We tried to regerate this library with gennfgt and genclntsh, but the problem remains the same. Does anyone already met this problem ? And then, what is the solution, how can we make the libclntsh.so compatible with the new kernel ?
Thank you for any help

If you're talking about the "so called" calendar Lov implemented using forms-objects (blocks etc.), you will have the problem with required items, for the the calendar will issue a GO_BLOCK which causes validation to happen, which will fail on the require item.
Solution?
If you're on 10G, theres also an example with a java-bean.
If on 6i, you could create a standalone form implementing the calendar-lov-functionality and do a call_form instead of a GO_BLOCK.

3.0.9.8.3 Issues on Tru64

Similar Messages

Maybe you are looking for