Cancel
Showing results for 
Search instead for 
Did you mean: 

Problem upgrading from Tc9.1 to 11.2

Solution Partner Creator Solution Partner Creator
Solution Partner Creator

We are attempting to upgrade from Tc 9.1 to 11.2 and running into an issue.  I was hoping someone might be able to offer some ideas.

 

Right now we are performing dry runs of the upgrade to make sure we have the documentation in order for actual production run.  We even successfully upgraded at one point.  However, when we performed the second dry run using the production Oracle server (with a new instance), the upgrade failed.  The issue occurs near the beginning of the upgrade process....

 

----------------------------------------------------------------------------
Executing command 6 of 250...
Section: pom-bootstrap-classes
File: D:\Siemens\Teamcenter11\install\foundation\upgrade_foundation_v20091.default
Command: install -mod_class -u=infodba -p=***** -g=dba ImanExportRecord POM_uninheritable_class +
----------------------------------------------------------------------------
Command start time: Wed Sep 19 11:11:42 AM 2018
Output:


Command end time: Wed Sep 19 11:11:42 AM 2018
Command duration: 0.609 seconds
ERROR: Execution of command 6 failed

 

Looking at the "install mod_class" syslog files, there are lots of errors but the first warning and error is as follows...

 

WARN  - 2018/09/19-16:11:42.690 UTC - NoId - 904 - OCIStmtExecute error -1 904 "SELECT ref_uid,vuid,process,type,aspace_uid FROM POM_LOCK WHERE ref_uid  = :1 AND aspace_uid=:2 ORDER BY ref_uid, aspace_uid" - Teamcenter.POM at D:\workdir\tc1140x_win64\src\foundation\pom\eim\eim_enquire.cxx(2303)
Arguments  :1 'qvCtOQtNIo1MyB'
           :2 'AAAAAAAAAAAAAA'

look for locks before saving
*** EIM_check_error: parse_error_offset 0

*** EIM_check_error: code -904
WARN  - 2018/09/19-16:11:42.690 UTC - NoId - 545001 - (EIM_trans_util) ORA-00904: "ASPACE_UID": invalid identifier
 - Teamcenter.POM at D:\workdir\tc1140x_win64\src\foundation\pom\eim\eim_trans_util.cxx(5337)


#######inside EIM_not_ok##########

sqlca_error_code is -904
EIM_db_error_code is 545001
ERROR - 2018/09/19-16:11:42.690 UTC - NoId - 545001
ERROR - 2018/09/19-16:11:42.690 UTC - NoId - *** EIM_check_error: 'look for locks before saving' code 545001 - Teamcenter at D:\workdir\tc1140x_win64\src\foundation\pom\eim\eim_trans_util.cxx(5049)

 

Any help is greatly appreciated.  Thanks!

 

6 REPLIES

Re: Problem upgrading from Tc9.1 to 11.2

Solution Partner Honored Contributor Solution Partner Honored Contributor
Solution Partner Honored Contributor
Can you run clearlocks -assert_all_dead before retrying the upgrade? It should have already ran as part of the upgrade though. You may have a table lock somewhere that is preventing Oracle from changing the data. What version of Oracle are you running?

Randy Ellsworth, Teamcenter Architect, Applied CAx, LLC
NX 11 | SW 2016 | Creo 4 | TcUA 11.4
Evaluating: AW 3.4

Re: Problem upgrading from Tc9.1 to 11.2

Solution Partner Creator Solution Partner Creator
Solution Partner Creator

I ran the clearlocks command from a CMD window and then retried the upgrade.  The same errors occured.

 

The database is Oracle 11gR2

 

The oracle db admin has looked at the current state of the database and said that is looks like the installer is trying to access a column that doesn't exist.  I don't understand why that would occur.  It's like the db is at 9.1 but the installer thinks it's already at 11.2.

 

 

Re: Problem upgrading from Tc9.1 to 11.2

Solution Partner Experimenter Solution Partner Experimenter
Solution Partner Experimenter

I also have this problem.Have you dealt with the problem?

Re: Problem upgrading from Tc9.1 to 11.2

Solution Partner Honored Contributor Solution Partner Honored Contributor
Solution Partner Honored Contributor
You can find the syslog for this failed shell command in TC_ROOT\logs\<subdir>\install*.syslog. Likely, it will contain useful infomation for debugging this error.

Randy Ellsworth, Teamcenter Architect, Applied CAx, LLC
NX 11 | SW 2016 | Creo 4 | TcUA 11.4
Evaluating: AW 3.4

Re: Problem upgrading from Tc9.1 to 11.2

Solution Partner Experimenter Solution Partner Experimenter
Solution Partner Experimenter

I've seen the syslog. It prompts the following error messages:

 

 update*.log

Executing ["install" -mod_class -u=infodba -p=***** -g=dba POM_stub POM_uninheritable_class +]
Output from command:  "install" -mod_class -u=infodba -p=***** -g=dba POM_stub POM_uninheritable_class +
Selected option: modify class properties.
Started POM v1000.0.1 ...
Failed to modify the schema - are there other users?: got 515145 error
Failed to modify the schema - are there other users?
Error 515145 from POM_modify_schema, meaning: POM_schema_locked
*** Caught an error while installing ... aborting!

 

install-mod_class*.log:

===pid AARtPZgnJsHkTB, module PM, process 00002b78:000003cc plmappsvr01, infodba logged in to plmappsvr01 on 2018 11 20 19:26 D:\Siemens\Teamcenter11\logs\flatten_class2b78ea40.syslog - Teamcenter.POM at D:\workdir\tc11301_a01_win64\src\foundation\pom\pom\pom_start.cxx(812)
ERROR - 2018/11/20-11:26:25.356 UTC - NoId - 515145
ERROR - 2018/11/20-11:26:25.356 UTC - NoId - There are other POM sessions active - unable to modify the schema.

 

But Error will not be displayed if i ran the command in CMD window.So I don't know how to dealt with.

 

I also ran the clearlocks,but nothing worked.

Re: Problem upgrading from Tc9.1 to 11.2

Solution Partner Creator Solution Partner Creator
Solution Partner Creator

I appologize for not updating this thread sooner but we were able to workaround the original problem by first adding more memory to the Oracle database server and second, performing the upgrade when the server was at its lowest usage rate.  Read on for a full description of what we found and how it was solved.

 

During the upgrade from 9.1 to 11.2, the upgrade gave an error that there was not enough memory on the Oracle database server.  Attempting the upgrade a second time (using the recover option) allowed the upgrade to continue without further memory errors but then we ran into other errors regarding the install -mod_class command. When we investigated this error, it appeared the database had attributes available that it shouldn't have.  I.E. The attributes were associated with 11.2 which hadn't been installed yet and this confused the installer.

 

What we believe happened was that some attributes were added to the database before the memory error occured, then when the recovery was attempted, these attributes shouldn't have existed and caused the errors.  This may be a bug in the installer but it's definitely a fringe case since the real issue was that the server didn't have enough memory to begin with.

 

The workaround was to revert the TC server back to 9.1 in a state that was prior to any failures, then ensure the Oracle database had enough memory before attempting the upgrade.  That way the initial upgrade could proceed completely.  To do this, we added memory to the Oracle server and since the Oracel database was contained on a shared server (which was used for other applications and other databases) we performed the upgrade at a time when the server was at its lowest usage rate (10PM Est in this case).  The upgrade was then successful.

 

-Rikh