Peoplesoft OVM PORTAL9.1

Few weeks ago a new Peoplesoft OVM template has been released, Portal Solution 9.1, announced here.
After HCM9.1, this is the second one, time to test it as well.

As usual, download from http://edelivery.oracle.com/linux (check the template 64bit : PeopleSoft VM Templates for PS 9.1 Release Media Pack for x86_64 (64 bit)).

Only two templates for this : database and App/Batch/PIA.

I’ll skip all the steps regarding Oracle VM Manager, Oracle VM Server configuration, uncompress and import of the template since this is the same exercise I did last year for HCM9.1, please if you want to know more, have a look to the tutorial on the top-right of the main blog page. I’ll start the description below from the starting of the new VM.

1. The database server
Select the VM and click on start
 
Then on console (be fast otherwise it will be configured as DHCP you certainly don’t want). 
 
Enter the password defined earlier when you created the template :

Give all what it needs, IP address and so on :

Here below, we observe the same CSS error as before for the HCM9.1 template, that fails, but it is not an issue for the next steps :


It continues, give the database name

Few notes :
1.1. Well, so far so good despite I choose automatic startup of the database (see screenshot above), the database does not start on the server boot (same as HCM9.1).
1.2. Note, the service is your_db_name.us.oracle.com :
[oracle@psovmpsdb ~]$ lsnrctl status

LSNRCTL for Linux: Version 11.1.0.7.0 - Production on 03-AUG-2010 13:04:20

Copyright (c) 1991, 2008, Oracle.  All rights reserved.

Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=psovmpsdb)(PORT=1521)))
STATUS of the LISTENER
------------------------
Alias                     LISTENER
Version                   TNSLSNR for Linux: Version 11.1.0.7.0 - Production
Start Date                03-AUG-2010 12:28:15
Uptime                    0 days 0 hr. 36 min. 5 sec
Trace Level               off
Security                  ON: Local OS Authentication
SNMP                      OFF
Listener Parameter File   /u01/app/oracle/product/11.1.0/db_1/network/admin/listener.ora
Listener Log File         /u01/app/oracle/product/11.1.0/db_1/log/diag/tnslsnr/psovmpsdb/listener/alert/log.xml
Listening Endpoints Summary...
  (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=psovmpsdb)(PORT=1521)))
  (DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC1521)))
Services Summary...
Service "P91TMPLT.us.oracle.com" has 1 instance(s).
  Instance "P91TMPLT", status READY, has 1 handler(s) for this service...
Service "P91TMPLT_XPT.us.oracle.com" has 1 instance(s).
  Instance "P91TMPLT", status READY, has 1 handler(s) for this service...
Service "XDB.us.oracle.com" has 1 instance(s).
  Instance "P91TMPLT", status READY, has 1 handler(s) for this service...
The command completed successfully
[oracle@psovmpsdb ~]$
1.3. SYSADM’s password is in upper case :
C:\>sqlplus sysadm/sysadm@p91tmplt

SQL*Plus: Release 11.1.0.7.0 - Production on Tue Aug 3 22:04:49 2010

Copyright (c) 1982, 2008, Oracle.  All rights reserved.

ERROR:
ORA-01017: invalid username/password; logon denied

Enter user-name:
C:\>sqlplus sysadm/SYSADM@p91tmplt

SQL*Plus: Release 11.1.0.7.0 - Production on Tue Aug 3 22:04:56 2010

Copyright (c) 1982, 2008, Oracle.  All rights reserved.
1.4. Language installed is only English :
SQL> select language_cd from pslanguages where installed=1;

LAN
---
ENG
1.5. The default profile of SYSADM and PEOPLE users expires password the 23-AUG-2010 :
SQL> select username,account_status,expiry_date,profile
  2  from   dba_users
  3  where username in ('PEOPLE','SYSADM');

USERNAME                       ACCOUNT_STATUS                   EXPIRY_DA PROFILE
------------------------------ -------------------------------- --------- ---------
PEOPLE                         OPEN                             23-AUG-10 DEFAULT
SYSADM                         OPEN                             23-AUG-10 DEFAULT

SQL> select *
  2  from dba_profiles
  3  where profile='DEFAULT'
  4  and limit is not null
  5  and limit ! ='UNLIMITED';

PROFILE                        RESOURCE_NAME                    RESOURCE LIMIT
------------------------------ -------------------------------- -------- ----------
DEFAULT                        FAILED_LOGIN_ATTEMPTS            PASSWORD 10
DEFAULT                        PASSWORD_VERIFY_FUNCTION         PASSWORD NULL
DEFAULT                        PASSWORD_GRACE_TIME              PASSWORD 7
DEFAULT                        PASSWORD_LIFE_TIME               PASSWORD 180
DEFAULT                        PASSWORD_LOCK_TIME               PASSWORD 1

You should change it to avoid future issues :
SQL> alter profile default limit
  2  failed_login_attempts unlimited
  3  password_life_time unlimited
  4  password_lock_time unlimited
  5  password_grace_time unlimited;

Profile altered.

SQL> select *
  2  from dba_profiles
  3  where profile='DEFAULT'
  4  and limit is not null
  5  and limit ! ='UNLIMITED';

PROFILE                        RESOURCE_NAME                    RESOURCE LIMIT
------------------------------ -------------------------------- -------- -----
DEFAULT                        PASSWORD_VERIFY_FUNCTION         PASSWORD NULL

2. The App/Batch/PIA server
Select the VM and click on start
 
Then on console (be fast otherwise it will be configured as DHCP you certainly don’t want). 
 
Enter the password defined earlier when you created the template :


 

Give all what it needs, IP address and so on :

   
Here it will continue if you want :

Well done, AppServer AND process scheduler start on the first shot :

It will continue by the Webserver if you want :



Done, we get the login page as expected :
 

Much appreciated, and really, except posting report, all works on the first shot ! Really well done !

The few errors revealed within HCM9.1 template has been solved (db connection, kernel settings…).
Now to be able to post report, you should do small effort on your own, create a report node repository and attached to the process server defined :

Note : the local node is already password configured (again one more improvement compared to the HCM9.1).
Everything should post correctly now :

There is one problem I encountered, I don’t know if this is Peoletools related (here it is 8.50.09) but it is rather annoying, and since I don’t have any other other env. in the same Peopletools release, I cannot reproduce somewhere else.
When I enter a text in any lookup field, I receive a Internet Explorer error :

Same issue in the role page of a user :

I’m sure this is not coming from template, but still….
Addendum : it seems to be an error introduced within Peopletools 8.50.07/08, here on OTN forum, a workaround to fix that issue (with TYPEAHD) given by Greg Kelly from Peopletools team. You could also have a look in MOS note #1147744.1 which describe the way to deactivate it.

Nicolas.

Oracle patchset 10.2.0.5 recut (Linux 32bit)

Important note for those who work on Linux 32bit, the patchset 10.2.0.5 has been recut and few fixes has been added to the original package dated from 30-APR-2010.
Do not forget to download it again to be sure to work on the latest version.
Read the note 1156958.1 for more information, this is only for 32bit version of Linux.

Nicolas.

Oracle Database Vault and Peoplesoft

If you want to secure and close down your database to all the non-authorized people, especially not seeing the data inside the database, what’s better than Database Vault ? Nothing.
And this is more and more right, especially on financial database or even on HR system hosting the salary and a lot of confidential information about the employees of the company. So, what’s refrain to use Database Vault ? Nothing. Almost nothing. Till 10gR2, Database Vault was an additional option to download and install onto the Oracle home directory, from 11g, it is included in the standard package, no need anymore additional software install.
For Peoplesoft however, you’ll need an additional component dedicated to Peoplesoft to create all the required policies.

Installation
I’m not going to describe all the installation steps, but there are several constraints when installing Database Vault, have a dedicated Oracle home, have EM Console, have one database (declared in /etc/oratab) to name few (note, at least in 10gR2 ASM is not supported by DBVault), but you can find everything in the well documented link.

Implementation
Once the Database Vault has been installed, you’ll need to apply it onto the database, following the installation through the dvca (Database Vault Creation Assistant).
Some new users are created to manage the database vault policies, owner and manager (you can decide to have only the owner). Only them can allow other users to do something on the database.

Peoplesoft package
Once your database has been “vaulted”, you have to install the dedicated Peoplesoft packages pointed above in the download page.

Maintenance
Whether the installation process might fail for several reasons, and it is not always easy to figure out why (the messages are not really clear), as long as you read very carefully the document, this is rather simple. As soon has the database is vaulted and Peoplesoft packages installed, no one else than SYSADM (the Peoplesoft objects’ owner) can see the data (even SYS/SYSTEM are “blind”). And more than that, SYSADM cannot connect through SQL*Plus (or any other non-Peoplesoft tools) to see data. Happy ? Well, yes and no.
In the documentation coming within the Peoplesoft package, it is clearly explained that a new user – PSFTDBA – has been created and authorized to do all the maintenance tasks in place of SYSADM through SQL*Plus for instance.
PSFTDBA can create/drop/alter objects in SYSADM’s with the ALTER SESSION SET SCHEMA=SYSADM. But again to be respectful to the vault policies, PSFTDBA is not able to select data from SYSADM. Very easy.

But for who know the Peoplesoft project build, that is not so simple. Most of the time, the “project build” generate script within CTAS (Create Table As Select) which obviously cannot work in that database vault context. So what ?
The only one known workaround is to ask to the database vault owner (or manager) to deactivate the PSFTDBA restriction, he can then select data from SYSADM, run the script and reactivate the restriction… It is rather against the rules, isn’t it ? If we have to disable the vault to apply project, nothing refrain to see data anymore.

I’d be curious if someone has already implemented Database Vault on a Peoplesoft database and how they are managing it. Feel free to comment out.

Enjoy it,

Nicolas.

About the PSU July 2010

Started from last year, Oracle delivered quarterly PSU (PatchSet Update) containing CPU as well as several one-off patches. So far, the PSUs were cumulatives, no need to apply any of the previous ones to be able to install the latest one, like for the vast majority of the patches and CPUs by the past. And that was prety nice.

But, the PSU of July 2010 is quite different. Indeed, it has a prerequisite. The PSUJUL2010 must be installed on top of the PSUAPR2010. For instance, PSUJUL2010 for 10gR2 is 10.2.0.4.5 needs PSUAPR2010 - 10.2.0.4.4. It is quite unusual and I find it rather silently changed.

That does not mean the next PSU (in OCT2010) will have necessarily the prerequisite of the current one -PSUJUL2010, it will depend of the type of PSU : merge patch or overlay.

It must be kept in mind since it is very sensitive, from now, you must pay an additioanl and particular attention about the type of PSU. But if you missed some PSUs and want to install a new one, that could implies additional work.

Find out more in the note 854428.1 (6. Overlay Patchset Updates)

Nicolas.

A GoldenGate database for Peoplesoft reports

I came across a very interesting MOS note, Deploying Oracle GoldenGate to Achieve Operational Reporting for PeopleSoft [ID 1114746.1]. It explains how build a GoldenGate (replication) database dedicated to Process Scheduler reports, instead of running reports on the primary database.
It takes full advantages of GoldenGate compared to a classic replication like standby database mechanism.
Indeed, GoldenGate allows us :
1) to replicate only a set of tables from source : who knows Peoplesoft database can realize the benefits in terms of space saving on secondary site.
2) to replicate in both directions, from primary to secondary site, but also from secondary to primary : this will be used to run a Process Scheduler on the secondary site, the tables containing the process submission will be replicated from primary to secondary, this will be checked by the Process Scheduler running on the secondary site and it will fire the process. Finally, as soon as the process is done on secondary site, the results are send back to the primary by the replication from secondary to the primary site.
Furthermore, some heavy reports can implies a lot of work regarding the queries and database tuning, and some specifics needs you would avoid on your primary database (bigger temporary tablespace, additional indexes…). So, if you have heavy report managed by the Process Scheduler to run, you could freed your database of this amount of reporting work.
Very well done, well explained in details (including what tables have to be replicated and/or excluded in what direction), this document is worth to read.

Nicolas.

Peoplesoft load without segments

As discussed few months ago here, Oracle 11gR2 offers the possibility to create table without segment. Particularly interesting on Peoplesoft database to drop down the size of the occupied size in the database, as I explained it is difficult to get rid off segment on existing table, but what happens on a new database load for a first Peopletools installation ?
Everything works as expected, after doing a database creation, just be sure to set the parameter deferred_segment_creation to TRUE on system level before the very first DataMover load :
SQL> show parameter deferred_segment_creation

NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
deferred_segment_creation            boolean     TRUE
SQL>
Then, after load, a quick check shows a lot of tables are empty (HCM9.1/Peopletools 8.50) :
SQL> select count(*) from user_tables where num_rows = 0;

  COUNT(*)
----------
     12101
But how many of them have a segment ?
SQL> select count(*)
  2  from   user_tables,dba_segments
  3  where  num_rows = 0
  4  and    segment_type='TABLE'
  5  and    table_name=segment_name;

  COUNT(*)
----------
         0
Yeah, actually none of the empty tables have segment !

Now to get rid of segment on existing table, what about export and import through DataMover ?
Just found a table with few rows, and truncate it (just for my test) :
SQL> truncate table PS_AA_COST_RT_JPN;

Table truncated.

SQL> select count(*) from dba_segments where segment_name='PS_AA_COST_RT_JPN' and segment_type='TABLE';

  COUNT(*)
----------
         1

Then with DMS, connect as PS and export the table :

Finally, import with REPLACE_ALL option :

Let’s check what happened on database side :
SQL> select count(*) from user_tables where table_name='PS_AA_COST_RT_JPN';

  COUNT(*)
----------
         1

SQL> select count(*) from dba_segments where segment_name='PS_AA_COST_RT_JPN' and segment_type='TABLE';

  COUNT(*)
----------
         0

Yeap, no more segment. So, finally, if you have an existing Peoplesoft database, and think about an upgrade to 11gR2, maybe a DataMover export and re-import of all the empty tables has to be considered (maybe that’s easier than “rebuild” everything through AppDesigner).

Nicolas.

CentOS 5.5 (and VMWare 2.0.2)

Yesterday the CentOS version 5.5 has been released. It is only few weeks after the Red Hat 5.5, this last update of CentOS was rather quick compared to the previous one.
So, time to try it.
I was especially looking for a solution for the issue I got within VMWare 2.0.2 on CentOS 5.4 as I described few months ago here.
So, before the upgrade I dropped the extra file I used to workaround the issue (please, read the article linked above), then after running a simple “yum list updates” and “yum update”, the system is upgraded :
[root@hercules /root]# rpm -q centos-release
centos-release-5-5.el5.centos
[root@hercules /root]#
I bounced the server, then run a “vmware-config.pl” from VMWare software binaries, and eventually everything is working fine. The system is stable, the VMWare webserver is not hanging and works fine. No need anymore the workaround needed on CentOS5.4 to downgrade some of the packages…
Thanks to the CentOS team for their hard work and to deliver this update so fast !

Nicolas.

Update (17-May-2010) : you may still have issues on some circumstances… please, read my comment below.

Oracle 11gR2 for Windows

About seven months after the Linux version of Oracle 11gR2, the version for Windows has been released last Friday.
So far, only the 64-bit edition is available, too bad for the 32-bit compliant only application, like Peoplesoft on Windows. Need to wait a little bit more for the 32-bit client on Windows. It still needs to have a previous version 32-bit, that’s rather annoying to have a discrepancy between client and server version. Well, the documentation for the 32-bit is already there, so it should come soon.

Anyway, for those who want to play with 11gR2 on Windows, here below you can download :
http://www.oracle.com/technology/software/products/database/index.html

Nicolas.

PS (06-APR-2010) : it was not that long to wait, the 32-bit version is now available from the same link above.

New blog entry tool

This is a new blog entry test done with a new (new for me) offline blog tool.

With a screenshot above.
All done with Windows Live Writer.

Just to test it. The most annoying thing with this is probably the new paragraph created when type “enter”, it is not a simple carriage return. The workaround is to modify the HTML code, remove the tags <p> and </p> and insert a <br> at the end of the line. But at least it is time and nerves saving compared with the too basic default one from Blogspot.

this is a test of code appearance
----------------------------------------------
Quick-configure menu -- domain: DMOHRMS9
----------------------------------------------
   Features                      Settings
  ==========                    ==========
1) Pub/Sub Servers   : Yes   15) DBNAME     :[DMOHRMS9]
2) Quick Server      : No    16) DBTYPE     :[ORACLE]
3) Query Servers     : No    17) UserId     :[PS]
4) Jolt              : Yes   18) UserPswd   :[PS]
5) Jolt Relay        : No    19) DomainID   :[DMOHRMS9]
6) WSL               : Yes   20) AddToPATH  :[/apps/oracle/product/11.1.0]
7) PC Debugger       : No    21) ConnectID  :[people]
8) Event Notification: No    22) ConnectPswd:[peop1e]
9) MCF Servers       : No    23) ServerName :[]
10) Perf Collator     : No    24) WSL Port   :[7000]
11) Analytic Servers  : No    25) JSL Port   :[9000]
12) Domains Gateway   : No    26) JRAD Port  :[9100]

PSAPPSRV.8363 (0) [04/04/10 16:12:23](1) GenMessageBox(200, 0, M): PS General SQL Routines: Missing or invalid version of SQL library libpsora (200,0)
PSAPPSRV.8363 (0) [04/04/10 16:12:23](1) GenMessageBox(0, 0, M): Database Signon: Could not sign on to database DMOHRMS9 with user PS.
PSAPPSRV.8363 (0) [04/04/10 16:12:23](0) Server failed to start

Nicolas.

Troubleshooting : process scheduler and distribution

When debugging process scheduler (or AppServer), we need to be sure about the certification matrix. It might appear basic advice, but it is very often forgotten. The certification matrix assure that the combination of component must work.
Note I said "must work", and not the other way around, if a combination is not certified, it does not necessarily mean it does not work, but at least if it is certified it must work.

The first two components to check are the OS wordsize and the database client version (32 or 64 bit). Depending of your Peopletools version, this could be very different, from a certification point of view as well as from a working point of view.
And this two points are not exclusive to the process scheduler, but are also valid for AppServer, that's why tests below could also contain example with AppServer as well as Prcs.

1. The OS wordsize
I don't have any troubleshoot tests cases, but I assume it must be easy to be respectful of this constraint.

Two situations Peopletools 8.49 :
=> Before Peopletools 8.49.14 : Linux 32bit only is certified
=> From Peopletools 8.49.14+ : Linux 64bit is also certified

Only one certification path with Peopletools 8.50 :
=> Peopletools 8.50.xx : 64bit only is certified (regardless the OS, Unix, Linux and Windows)

2. The database client libraries
2.1 Peopletools 8.49 (and below)
Whatever the wordsize of the OS the Peopletools are running on, it must use the Oracle 32bit libraries.
Here we can distinguished two cases, before Oracle 11gR2 and from 11gR2.
2.1.1 Before 11gR2
Here a test case with Peopletools 8.49.20 working with Oracle 11.1.0.7 64-bit on Linux OEL 5.3 64-bit. The test below is being with AppServer, but this is exactly the same with Process Scheduler.

[oracle@orion2:/apps/oracle/product/11.1.0/bin]$ file oracle
oracle: setuid setgid ELF 64-bit LSB executable, AMD x86-64, version 1 (SYSV), for GNU/Linux 2.6.9, dynamically linked (uses shared libs), for GNU/Linux 2.6.9, not stripped
[oracle@orion2:/apps/oracle/product/11.1.0/bin]$
----------------------------------------------
Quick-configure menu -- domain: DMOHRMS9
----------------------------------------------
   Features                      Settings
  ==========                    ==========
1) Pub/Sub Servers   : Yes   15) DBNAME     :[DMOHRMS9]
2) Quick Server      : No    16) DBTYPE     :[ORACLE]
3) Query Servers     : No    17) UserId     :[PS]
4) Jolt              : Yes   18) UserPswd   :[PS]
5) Jolt Relay        : No    19) DomainID   :[DMOHRMS9]
6) WSL               : Yes   20) AddToPATH  :[/apps/oracle/product/11.1.0]
7) PC Debugger       : No    21) ConnectID  :[people]
8) Event Notification: No    22) ConnectPswd:[peop1e]
9) MCF Servers       : No    23) ServerName :[]
10) Perf Collator     : No    24) WSL Port   :[7000]
11) Analytic Servers  : No    25) JSL Port   :[9000]
12) Domains Gateway   : No    26) JRAD Port  :[9100]
...
Booting server processes ...

exec PSWATCHSRV -A -- -ID 114790 -C psappsrv.cfg -D DMOHRMS9 -S PSWATCHSRV :
     process id=8362 ... Started.
exec PSAPPSRV -s@../psappsrv.lst -s@../psqcksrv.lst -sICQuery -sSqlQuery:SqlRequest -- -C psappsrv.cfg -D DMOHRMS9 -S PSAPPSRV :
     CMDTUX_CAT:1685: ERROR: Application initialization failure

tmboot: CMDTUX_CAT:827: ERROR: Fatal error encountered; initiating user error handler

exec tmshutdown -qy

==============ERROR!================
Boot attempt encountered errors!. Check the TUXEDO log for details.
==============ERROR!================

Do you wish to see the error messages in the APPSRV.LOG file? (y/n) [n] :y

PSADMIN.8321 (0) [04/04/10 16:12:14](0) Begin boot attempt on domain DMOHRMS9
PSWATCHSRV.8362 (0) [04/04/10 16:12:22] Checking process status every 120 seconds
PSWATCHSRV.8362 (0) [04/04/10 16:12:22] Server started
PSAPPSRV.8363 (0) [04/04/10 16:12:22](0) PeopleTools Release 8.49.20 (Linux) starting
PSAPPSRV.8363 (0) [04/04/10 16:12:23](0) Cache Directory being used: /apps/psoft/hrms9/appserv/DMOHRMS9/CACHE/PSAPPSRV_1/
PSAPPSRV.8363 (0) [04/04/10 16:12:23](1) GenMessageBox(200, 0, M): PS General SQL Routines: Missing or invalid version of SQL library libpsora (200,0)
PSAPPSRV.8363 (0) [04/04/10 16:12:23](1) GenMessageBox(0, 0, M): Database Signon: Could not sign on to database DMOHRMS9 with user PS.
PSAPPSRV.8363 (0) [04/04/10 16:12:23](0) Server failed to start
PSWATCHSRV.8362 (0) [04/04/10 16:12:24] Shutting down
PSADMIN.8321 (0) [04/04/10 16:12:30](0) End boot attempt on domain DMOHRMS9

Do you wish to see the error messages in the TUXLOG.040410 file? (y/n) [n] :

We can check the stderr file from the domain we are trying to start :

[hrms9@orion2:/apps/psoft/hrms9/appserv/DMOHRMS9]$ more stderr
dlopen in libpscompat failed for 'libpsora.so': libclntsh.so.9.0: cannot open shared object file: No such file or directory

As suggested in the certification matrix on My Oracle Support, a symbolic link must be created to workaround this. The Oracle is a 64-bit, but we can use the lib32 libraries :

[oracle@orion2:/apps/oracle/product/11.1.0/lib]$ ln -s /apps/oracle/product/11.1.0/lib32/libclntsh.so.11.1 libclntsh.so.9.0

And restart the AppServer works fine :

Booting server processes ...

exec PSWATCHSRV -A -- -ID 114790 -C psappsrv.cfg -D DMOHRMS9 -S PSWATCHSRV :
     process id=8460 ... Started.
exec PSAPPSRV -s@../psappsrv.lst -s@../psqcksrv.lst -sICQuery -sSqlQuery:SqlRequest -- -C psappsrv.cfg -D DMOHRMS9 -S PSAPPSRV :
     process id=8462 ... Started
....
15 processes started.

2.1.2 With 11gR2
Here a test case with Peopletools 8.49.20 working with Oracle 11.2.0.1 64-bit on Linux OEL 5.3 64-bit.

Booting admin processes ...

exec BBL -A :
     process id=8992 ... Started.

Booting server processes ...

exec PSMSTPRC -A -- -C psprcs.cfg -CD DMOHRMS9 -PS PSUNX -A start -S PSMSTPRC :
     CMDTUX_CAT:1685: ERROR: Application initialization failure

tmboot: CMDTUX_CAT:827: ERROR: Fatal error encountered; initiating user error handler

exec tmshutdown -y

Shutting down all admin and server processes in /apps/psoft/hrms9/appserv/prcs/DMOHRMS9/PSTUXCFG

Shutting down server processes ...


Shutting down admin processes ...

     Server Id = 0 Group Id = orion2.phoenix-nga Machine = orion2.phoenix-nga:       shutdown succeeded
1 process stopped.

And the stderr file :

[hrms9@orion2:/apps/psoft/hrms9/appserv/prcs/DMOHRMS9]$ more stderr
dlopen in libpscompat failed for 'libpsora.so': libclntsh.so.9.0: cannot open shared object file: No such file or directory

Within Oracle 11gR2 64-bit, there is no 32-bit libraries (as I described here), so we can try the symbolic link as above, but it will not work :

[oracle@orion2:/apps/oracle/product/11.2.0/lib]$ ln -s libclntsh.so.11.1 libclntsh.so.9.0
...
Booting server processes ...

exec PSWATCHSRV -A -- -ID 111990 -C psappsrv.cfg -D DMOHRMS9 -S PSWATCHSRV :
     process id=7625 ... Started.
exec PSAPPSRV -s@../psappsrv.lst -s@../psqcksrv.lst -sICQuery -sSqlQuery:SqlRequest -- -C psappsrv.cfg -D DMOHRMS9 -S PSAPPSRV :
     CMDTUX_CAT:1685: ERROR: Application initialization failure

tmboot: CMDTUX_CAT:827: ERROR: Fatal error encountered; initiating user error handler

exec tmshutdown -qy

==============ERROR!================
Boot attempt encountered errors!. Check the TUXEDO log for details.
==============ERROR!================

Do you wish to see the error messages in the APPSRV.LOG file? (y/n) [n] :y

PSADMIN.7615 (0) [04/04/10 16:00:30](0) Begin boot attempt on domain DMOHRMS9
PSWATCHSRV.7625 (0) [04/04/10 16:00:38] Checking process status every 120 seconds
PSWATCHSRV.7625 (0) [04/04/10 16:00:38] Server started
PSAPPSRV.7626 (0) [04/04/10 16:00:39](0) PeopleTools Release 8.49.20 (Linux) starting
PSAPPSRV.7626 (0) [04/04/10 16:00:39](0) Cache Directory being used: /apps/psoft/hrms9/appserv/DMOHRMS9/CACHE/PSAPPSRV_1/
PSAPPSRV.7626 (0) [04/04/10 16:00:39](1) GenMessageBox(200, 0, M): PS General SQL Routines: Missing or invalid version of SQL library libpsora (200,0)
PSAPPSRV.7626 (0) [04/04/10 16:00:39](1) GenMessageBox(0, 0, M): Database Signon: Could not sign on to database DMOHRMS9 with user PS.
PSAPPSRV.7626 (0) [04/04/10 16:00:39](0) Server failed to start
PSWATCHSRV.7625 (0) [04/04/10 16:00:40] Shutting down
PSADMIN.7615 (0) [04/04/10 16:00:46](0) End boot attempt on domain DMOHRMS9

Do you wish to see the error messages in the TUXLOG.040410 file? (y/n) [n] :
...
[hrms9@orion2:/apps/psoft/hrms9/appserv/DMOHRMS9]$ more stderr
dlopen in libpscompat failed for 'libpsora.so': libclntsh.so.9.0: wrong ELF class: ELFCLASS64

As expected, that does not work, the 64-bit libraries are not accepted by Peopletools 8.49.xx.
A real client 32-bit must be installed, and linked to the AppServer and Process Scheduler.
After that install, reconfigure the domain, everything is started :

----------------------------------------------
Quick-configure menu -- domain: DMOHRMS9
----------------------------------------------
   Features                      Settings
  ==========                    ==========
1) Pub/Sub Servers   : Yes   15) DBNAME     :[DMOHRMS9]
2) Quick Server      : No    16) DBTYPE     :[ORACLE]
3) Query Servers     : No    17) UserId     :[PS]
4) Jolt              : Yes   18) UserPswd   :[PS]
5) Jolt Relay        : No    19) DomainID   :[DMOHRMS9]
6) WSL               : Yes   20) AddToPATH  :[/apps/oracle/product/11.2.0_client_32bit]
7) PC Debugger       : No    21) ConnectID  :[people]
8) Event Notification: No    22) ConnectPswd:[peop1e]
9) MCF Servers       : No    23) ServerName :[]
10) Perf Collator     : No    24) WSL Port   :[7000]
11) Analytic Servers  : No    25) JSL Port   :[9000]
12) Domains Gateway   : No    26) JRAD Port  :[9100]

   Actions
  =========
13) Load config as shown
14) Custom configuration
h) Help for this menu
q) Return to previous menu

HINT: Enter 15 to edit DBNAME, then 13 to load

Enter selection (1-26, h, or q):
Booting server processes ...

exec PSWATCHSRV -A -- -ID 184396 -C psappsrv.cfg -D DMOHRMS9 -S PSWATCHSRV :
     process id=7907 ... Started.
exec PSAPPSRV -s@../psappsrv.lst -s@../psqcksrv.lst -sICQuery -sSqlQuery:SqlRequest -- -C psappsrv.cfg -D DMOHRMS9 -S PSAPPSRV :
     process id=7908 ... Started.
...
15 processes started.

2.2 Peopletools 8.50
2.2.3 Linux
On Linux, Peopletools 8.50 is fully certified to run on Oracle 64-bit client. At least on Linux, there is no issue anymore, 32 or 64-bit client should be ok. No need anymore the symbolic link workaround described above for previous Peopletools version, and no extra client need.

2.2.2 Windows
On Windows, however, it is a bit confusing. Once again, Peopletools 8.50 (Apps/Batch) are certified only on 64-bit OS server. But, the database client must be a 32-bit.
Let's see.
=> Oracle client 64bit libraries
Here a test with Peopletools 8.50.02, start process scheduler on W2k8 64bit, and Oracle 10.2.0.4 64-bit.

Booting admin processes ...

exec BBL -A :
     process id=2448 ... Started.

Booting server processes ...

exec PSMSTPRC -o ".\LOGS\stdout" -e ".\LOGS\stderr" -A -- -CD h92tmplt -PS PSNT -A start -S PSMSTPRC :
     CMDTUX_CAT:1685: ERROR: Application initialization failure
tmboot: CMDTUX_CAT:827: ERROR: Fatal error encountered; initiating user error handler

tmshutdown -y
Shutting down all admin and server processes in F:\appl\pt85002\appserv\prcs\h92tmplt\PSTUXCFG
Shutting down server processes ...

Shutting down admin processes ...
     Server Id = 0 Group Id = ANTLIA Machine = ANTLIA:       shutdown succeeded
1 process stopped.

As expected, that does not start and as usual, first file to check, stderr :

LoadLibraryA() in pscompat.dll failed for 'PSORA.dll': reason=126

Well, not very helpful message, but seeing closer the configuration, the given DBBIN did not includ the \bin directory. It should, let's change it :

Do you want to change any values (y/n)? [n]:y
 PrcsServerName [PSNT]:
 DBBIN [F:\appl\oracle\10.2.0.4\server_64bit]:F:\appl\oracle\10.2.0.4\server_64bit\bin
 Max Reconnect Attempt [12]:

New error on prcs start, stderr file is now :

LoadLibraryA() in pscompat.dll failed for 'PSORA.dll': reason=193

Not helpful neither to define that's a client wordsize error...

=> Oracle client 32bit libraries
Let's change now the client to a 32-vit one installed onto the same server :

...
Do you want to change any values (y/n)? [n]:y
 PrcsServerName [PSNT]:
 DBBIN [F:\appl\oracle\10.2.0.4\server_64bit\BIN]:F:\appl\oracle\10.2.0.4\client_32bit\BIN
 Max Reconnect Attempt [12]:
 Reconnection Interval [300]:
 ...

And boot start process scheduler :

Booting admin processes ...

exec BBL -A :
     process id=1152 ... Started.

Booting server processes ...

exec PSMSTPRC -o ".\LOGS\stdout" -e ".\LOGS\stderr" -A -- -CD h91tmplt -PS PSNT -A start -S PSMSTPRC :
     process id=2476 ... Started.
exec PSAESRV -o ".\LOGS\stdout" -e ".\LOGS\stderr" -- -CD h91tmplt -S PSAESRV :
     process id=2244 ... Started.
...
8 processes started.

Ok, it is started now.

This first part could be very disapointing sometimes and time spending, until to be very cautious before starting a Peoplesoft installation.

Note : all the tests below have been done from a W2k8 64bit workstation, on Peopletools 8.50.02 , Peoplesoft OVM database server (Peopletools 8.50.02/HCM9.1), Peoplesoft OVM App/Batch server (Peopletools 8.50.02) and Peoplesoft OVM PIA server (Peopletools 8.50.02).

3. The user to be used
A very basic user as been created onto the database (through the front end application, no role, no permission).
And a new process scheduler created and configured to be started with that user.
The prcs failed to start :

------------------------------------------------------------
Quick-configure menu -- Scheduler for Database: h91tmplt
------------------------------------------------------------
   Features                      Settings
  ==========                    ==========
1) Master Schdlr  : Yes     5) DBNAME     :[h91tmplt]
2) App Eng Server : Yes     6) DBTYPE     :[ORACLE]
                           7) PrcsServer :[PSUNX]
                           8) UserId     :[user_prcs]
                           9) UserPswd   :[user_prcs]
                          10) ConnectID  :[people]
                          11) ConnectPswd:[peop1e]
                          12) ServerName :[]
   Actions                13) Log/Output Dir:[%PS_SERVDIR%/log_output]
  =========               14) SQRBIN     :[%PS_HOME%/bin/sqr/%PS_DB%/bin]
3) Load config as shown    15) AddToPATH  :[%PS_HOME%/cblbin]
4) Custom configuration
h) Help for this menu
q) Return to previous menu
...
Starting Process Scheduler Server PSUNX for Database h91tmplt ...

Booting all admin and server processes in /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/PSTUXCFG
INFO: Oracle Tuxedo, Version 10.3.0.0, 64-bit, Patch Level (none)

Booting admin processes ...

exec BBL -A :
     process id=1970 ... Started.

Booting server processes ...

exec PSMSTPRC -o ./LOGS/stdout -e ./LOGS/stderr -A -- -CD h91tmplt -PS PSUNX -A start -S PSMSTPRC :
     CMDTUX_CAT:1685: ERROR: Application initialization failure

tmboot: CMDTUX_CAT:827: ERROR: Fatal error encountered; initiating user error handler

exec tmshutdown -y

Shutting down all admin and server processes in /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/PSTUXCFG

Shutting down server processes ...

Shutting down admin processes ...

     Server Id = 0 Group Id = psovmab.phoenix.nga Machine = psovmab.phoenix.nga:     shutdown succeeded
1 process stopped.

Unfortunately, in that case either, nothing helpful from the log files :

[psadm2@psovmab ps]$ cd pt/8.50/appserv/prcs/h91tmplt/
[psadm2@psovmab h91tmplt]$ ls
Archive  CACHE  LOGS  PSTUXCFG  files  log_output  psprcs.cfg  psprcsrv.cfx  psprcsrv.env  psprcsrv.ubb  psprcsrv.ubx  psprcsrv.val  psstat.in
[psadm2@psovmab h91tmplt]$ cd LOGS
[psadm2@psovmab LOGS]$ ls -lrt
total 24
-rw-r--r-- 1 psadm2 oracle    0 Apr  4 15:44 stdout
-rw-r--r-- 1 psadm2 oracle    0 Apr  4 15:44 stderr
-rw-r--r-- 1 psadm2 oracle  315 Apr  4 15:44 MSTRSCHDLR_0404.LOG
-rw-r--r-- 1 psadm2 oracle 1369 Apr  4 15:44 TUXLOG.040410
[psadm2@psovmab LOGS]$ more TUXLOG.040410
154435.psovmab.phoenix.nga!tmloadcf.1966.3461187040.-2: 04-04-2010: client high water (0), total client (0)
154435.psovmab.phoenix.nga!tmloadcf.1966.3461187040.-2: 04-04-2010: Tuxedo Version 10.3.0.0, 64-bit
154435.psovmab.phoenix.nga!tmloadcf.1966.3461187040.-2: CMDTUX_CAT:879: INFO: A new file system has been created. (size = 980 512-byte blocks)
154435.psovmab.phoenix.nga!tmloadcf.1966.3461187040.-2: CMDTUX_CAT:871: INFO: TUXCONFIG file /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/PSTUXCFG has been created
154443.psovmab.phoenix.nga!BBL.1970.2525654112.0: 04-04-2010: Tuxedo Version 10.3.0.0, 64-bit, Patch Level (none)
154443.psovmab.phoenix.nga!BBL.1970.2525654112.0: LIBTUX_CAT:262: INFO: Standard main starting
154443.psovmab.phoenix.nga!PSMSTPRC.1971.4055057680.-2: 04-04-2010: Tuxedo Version 10.3.0.0, 64-bit
154443.psovmab.phoenix.nga!PSMSTPRC.1971.4055057680.-2: LIBTUX_CAT:262: INFO: Standard main starting
154444.psovmab.phoenix.nga!PSMSTPRC.1971.4055057680.-2: LIBTUX_CAT:250: ERROR: tpsvrinit() failed
154444.psovmab.phoenix.nga!tmboot.1969.236119584.-2: 04-04-2010: Tuxedo Version 10.3.0.0, 64-bit
154444.psovmab.phoenix.nga!tmboot.1969.236119584.-2: tmboot: CMDTUX_CAT:827: ERROR: Fatal error encountered; initiating user error handler
154447.psovmab.phoenix.nga!BBL.1970.2525654112.0: CMDTUX_CAT:26: INFO: The BBL is exiting system
[psadm2@psovmab LOGS]$ more MSTRSCHDLR_0404.LOG
PSMSTPRC.1971 (0) [04/04/10 15:44:43](0) PeopleTools Release 8.50.02 (Linux) starting. Tuxedo server is BASE(1)/102
PSMSTPRC.1971 (0) [04/04/10 15:44:43](0) Cache Directory being used: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/CACHE/PSMSTPRC_102/
PSMSTPRC.1971 (0) [04/04/10 15:44:44](0) Server failed to start

And set some trace level does not help neither. There is a missing ROLE for that user user_prcs defined to managed the prcs : AppServer Administrator
Once this ROLE is added to the user_prcs user, the process scheduler is able to boot.

From this point, it is assuming the process scheduler is started. The following will be about the possible distribution issues.

4. The report node definition
Since the process scheduler is booted, let's run a process (AEMINITEST is the simplest one).
It won't post, here message logs :
And the prcs logs files :

[psadm2@psovmab LOGS]$ tail -20 stdout
     Process Type: Application Engine
=====================================================================
OprId = user_prcs
PS_TOKEN=qQAAAAQDAgEBAAAAvAIAAAAAAAAsAAAABABTaGRyAk4Acwg4AC4AMQAwABTz9QlIS4gf8bE/nzl4yBgXufOPHWkAAAAFAFNkYXRhXXicTYo9DkBQEIS/hygV7uGFRziBn0oEvUL0QtzO4QyVTb6Z2Z09gcD3jJHfHt/EFycbBwu7dNUW1vS0RAMTDbOajpHCkZKJRBQ/LbHkyhan3lLxfla6OHgAOKYNfw==

NULL HTTP response - check Report Repository web server. (63,70)
OprId = user_prcs
PS_TOKEN=pwAAAAQDAgEBAAAAvAIAAAAAAAAsAAAABABTaGRyAk4Acwg4AC4AMQAwABShoTHXH8OI+PFN7JZBHEtlDZDrAmcAAAAFAFNkYXRhW3icTYo7DkBQFETPQ5QK+yA8vxX4VCLoFaIXYncWZ6jcZM7M5M4JeK5jjPx2+C68ONk4WNjFVc2v6WkJBiYaZn06RnJLQipFUv5jSUym/LIQK6x29tvyADjODXs=

NULL HTTP response - check Report Repository web server. (63,70)
OprId = user_prcs
PS_TOKEN=qAAAAAQDAgEBAAAAvAIAAAAAAAAsAAAABABTaGRyAk4Acwg4AC4AMQAwABQP+W2bMOEOBikcFsPqiXpcc8FbrGgAAAAFAFNkYXRhXHicTYo7DkBQEEWPT5QK+/Di87ACn0oEvUL0QuzO4lwqk5z7mZkT8D3XceS3yzfRxcnGwcIuXdWCmp6WcGCiYdalY8RmJKQiFvanJYZc2VCoGSoy/b0bCw84wQ2A

NULL HTTP response - check Report Repository web server. (63,70)
=================================Error===============================
Message:       Unable to post report/log file for Process Instance: 891, Report Id: 737
       Process Name: AEMINITEST, Type: Application Engine
       Description: Simple AE test program
       Directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
=====================================================================
[psadm2@psovmab LOGS]$ tail -20 DSTAGNT_0404.LOG
PSDSTSRV.2600 (1) [04/04/10 16:34:22 PostReport](3)     HTTP transfer error.
PSDSTSRV.2600 (1) [04/04/10 16:34:22 PostReport](3)      Post Report Elapsed Time: 1.6900
PSDSTSRV.2600 (2) [04/04/10 16:34:35 PostReport](3) Number of new entries to process: 1
PSDSTSRV.2600 (2) [04/04/10 16:34:35 PostReport](3)   1.  Process Instance: 891/Report Id: 737/Descr: Simple AE test program
PSDSTSRV.2600 (2) [04/04/10 16:34:35 PostReport](3)                      from directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
PSDSTSRV.2600 (2) [04/04/10 16:34:35 PostReport](1) PSJNI: Java exception thrown: java.io.IOException: Stream closed.
PSDSTSRV.2600 (2) [04/04/10 16:34:35 PostReport](3)     HTTP transfer error.
PSDSTSRV.2600 (2) [04/04/10 16:34:35 PostReport](3)      Post Report Elapsed Time: 0.0100
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](3) Number of new entries to process: 1
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](3)   1.  Process Instance: 891/Report Id: 737/Descr: Simple AE test program
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](3)                      from directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](1) PSJNI: Java exception thrown: java.io.IOException: Stream closed.
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](3)     HTTP transfer error.
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](3)      Post Report Elapsed Time: 0.0200
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](1) =================================Error===============================
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](1)      Unable to post report/log file for Process Instance: 891, Report Id: 737
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](2)           Process Name: AEMINITEST, Type: Application Engine
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](2)           Description: Simple AE test program
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](2)           Directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
PSDSTSRV.2600 (3) [04/04/10 16:34:50 PostReport](2) =====================================================================
[psadm2@psovmab LOGS]$

Ok, the Report node is missing, let's add one.
But after bouncing the prcs, re-send the content does not help. Next step.

5. The local node definition
Now the error is a littel bit different :

[psadm2@psovmab LOGS]$ tail -20 stdout

The XML file returned by the web server is invalid. (63,94)

XML document object creation failed. (63,102)

Unable to process HTTP reply from Report Repository. (63,73)
OprId = user_prcs
PS_TOKEN=pwAAAAQDAgEBAAAAvAIAAAAAAAAsAAAABABTaGRyAk4Acwg4AC4AMQAwABRNcCgBNFm41dbNd6td8xOUAsInWGcAAAAFAFNkYXRhW3icTYk7DkBAAESfT5QK9yAsNnEAn0oEvUL0YuN2DmeiMsm8ycw4IAx8z1M+Pp+SG8fBxcYp7mpRy0hPPLHQseoZmKkMOYWcytWPlowa87EWDaXWRmnhBTkBDYY=

The XML file returned by the web server is invalid. (63,94)

XML document object creation failed. (63,102)

Unable to process HTTP reply from Report Repository. (63,73)
=================================Error===============================
 Message:       Unable to post report/log file for Process Instance: 891, Report Id: 737
          Process Name: AEMINITEST, Type: Application Engine
          Description: Simple AE test program
          Directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
=====================================================================
[psadm2@psovmab LOGS]$ tail -20 DSTAGNT_0404.LOG
PSDSTSRV.2832 (1) [04/04/10 16:52:24 PostReport](3) PSJNI: Created a Java VM instance
PSDSTSRV.2832 (1) [04/04/10 16:52:24 PostReport](3) PSJNI: Set Context Class Loader
PSDSTSRV.2832 (1) [04/04/10 16:52:25 PostReport](3)     HTTP transfer error.
PSDSTSRV.2832 (1) [04/04/10 16:52:25 PostReport](3)      Post Report Elapsed Time: 0.8700
PSDSTSRV.2832 (2) [04/04/10 16:52:39 PostReport](3) Number of new entries to process: 1
PSDSTSRV.2832 (2) [04/04/10 16:52:39 PostReport](3)   1.  Process Instance: 891/Report Id: 737/Descr: Simple AE test program
PSDSTSRV.2832 (2) [04/04/10 16:52:39 PostReport](3)                      from directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
PSDSTSRV.2832 (2) [04/04/10 16:52:39 PostReport](3)     HTTP transfer error.
PSDSTSRV.2832 (2) [04/04/10 16:52:39 PostReport](3)      Post Report Elapsed Time: 0.1000
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](3) Number of new entries to process: 1
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](3)   1.  Process Instance: 891/Report Id: 737/Descr: Simple AE test program
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](3)                      from directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](3)     HTTP transfer error.
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](3)      Post Report Elapsed Time: 0.0900
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](1) =================================Error===============================
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](1)      Unable to post report/log file for Process Instance: 891, Report Id: 737
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](2)           Process Name: AEMINITEST, Type: Application Engine
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](2)           Description: Simple AE test program
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](2)           Directory: /home/psadm2/ps/pt/8.50/appserv/prcs/h91tmplt/log_output/AE_AEMINITEST_891
PSDSTSRV.2832 (3) [04/04/10 16:52:54 PostReport](2) =====================================================================
[psadm2@psovmab LOGS]$

The local node must be trusted
And password authentication must be set.
After this setting and bounce the prcs, an other error to the next step.

6. The missing role - ProcessSchedulerAdmin
Now the error is the following :

OprId = user_prcs
PS_TOKEN=pgAAAAQDAgEBAAAAvAIAAAAAAAAsAAAABABTaGRyAk4Acwg4AC4AMQAwABQ69uNPo1aucPuKHZaK64OrsLUWgmYAAAAFAFNkYXRhWnicTYk7DkBQFETPQ5QK+yD+nwX4VCLoFaIXYncWZ6Jyk3NmJvcCHNsyRvlYfOffXOycrBzypuU2DHR4IzMtiz49E1lCRCwCkf1cEsqRnH+9IKWiFjG8OSANjw==

Access Denied, unable to post file(s) to repository. (63,124)

SchedulerTransfer Servlet error. (63,74)

Looks like we are close to the goal, but a right is missing, the role to post the report.
Let's add the ProcessSchedulerAdmin role to the user user_prcs who manage the process scheduler :
And bounce the prcs.
Now it is posting.


Of course, it is non-exhaustive testing, there are many other possible cases. But hopefully it is cover a large part and enough to debug most of the process scheduler issues.

Nicolas.