Failed to start CSS - PeopleSoft OVM Templates

Hi All!
I am trying to configure PS DB image and it ends up with 'Failed to start CSS' message. After that, I cleaned the image using '/opt/oracle/psft/vm/oraclevm-template' and increased the timeout in '/u01/dbreconfig.sh' script.
And still no success. Can anyone direct me how to get rid of this? Thanks! 

Yes, right, I even didn't notice it, but here the same, see my screenshot :
http://1.bp.blogspot.com/_GAPqu9CUk1k/SvXjoNudhgI/AAAAAAAABC4/Isc3cIHR7Jw/s1600-h/PSOVM_VMWARE_014.jpg
Despite the message, the database is created well, but unable to auto-start, see my post here :
http://gasparotto.blogspot.com/2009/11/peoplesoft-vm-about-database-image.html
I'll add this issue in the thread I created onto Oracle VM Virtualization forum :
Peoplesoft VM template
Nicolas. 

Oh, yeah. Thank you!
The problem is that in my case database won't start at all. I will double check everything again. 

You may have to start it by yourself.
Did you check if the folder with your DBNAME has been created somewhere under /opt/oracle/psft/ptdb/... ? In my case the db was created, I can connect also.
Nicolas. 

The folder is there, actually.But the next message after 'Failed to start CSS' is 'Failed to install database', something like that :)
How should I start the database? svrmgr is not there and my knowledge of oracle on linux is very poor... 

I as said earlier, despite the message*s* the db is created.
To start it, you can do :
su - oracle
lsnrctl start
export ORACLE_SID=YOUR_DBNAME
sqlplus / as sysdba
startupNicolas.
add the connection to oracle OS user
Edited by: N. Gasparotto on Nov 10, 2009 3:26 PM 

Thank you, now DB is started. 

"Failed to start CSS" followed by "Failed to install Oracle database" is a known issue and I believe we have documented it. We are working with the base DB template folks to identify the issue. These messages are coming from the base oracle DB template. The oracle database that is failing to install is the default oracle database 'orcl' and not a PeopleSoft database.
Once you enter any key, you will be prompted for PeopleSoft Database Name. This database will be initialized and started automatically during the boot process. Can you please verify if this is the case.
We are working with the base DB template folks to understand the CSS issue and hopefully get a resolution.
-Sheshi
(Tools) 

Thank you for you feedback.
Yes it is documented in the [pdf doc|http://download.oracle.com/docs/cd/E16146_01/psft/acrobat/ovmhcm91_1009.pdf]
"+Note. You may see error messages about the Oracle database after responding to this prompt. In addition,+
+the last line that asks for the local bin directory may not appear. You can ignore the error messages, or the+
+missing query, and continue+"
But it is not quite clear that it's regarding CSS.
And that's right the database is starting on the boot, but only on the first boot, not on the following ones.
The oracle database that is failing to install is the default oracle database 'orcl' and not a PeopleSoft database.And it is visible because in the /etc/oratab this database is declared, not the one we created. That could explain why the database is not started automatically on server boot.
Nicolas.
Edited by: N. Gasparotto (bis) on Nov 20, 2009 11:08 PM

Related

oms10g: Many core-files

Hi all!
Hi hope someone could help me.
We use 10g Enterprise Grid Control and run our backups over this tool.
Today we get a lot of mails that the agent of the clients could not upload their data.
So I found the similar problem: The space is full.
Under the directory oms10g I found 15000 files called for example core.1020.
May I could delete some of this files because they need 30G space...
Thank you for your answer
Kind Regards
René 
do you know why its core dumping? 
Hi,
no I don't know. I restart the server yesterday because the server runs since 1 year without reboot.
And today I get some error-mails like
Message=Count of targets not uploading exceeded the critical threshold (0). Current value: 3#
Message=Agent has stopped monitoring. The following errors are reported : COLL_DISABLED|DISK_FULL.
So you mean this are dump-files so where can I change the parameter? Maybe my ex-colleague changed a parameter and have forgotten to restart the machine and it is now activated since I have rebooted the machine.
Kind Regards
René
EDIT: Ok - so I am sure! After reboot the core-dumping would be activated. The grid control already creates dumpfiles without a content. So where can I find the parameter for dumping?
As I have written: Maybe my ex-colleague changed the parameter and will be enabled after reboot (restart the database). Is this mind correct?
Edited by: pairosilva on Oct 28, 2008 2:10 PM 
It is likely that the SYS password was changed when you ex-colleague left. If this is the case you will need to reset the value in targets.xml to the current password and restart OEM. Do a search on targets.xml to see what else needs to be changed if this is the case. 
Hmmm... Do you really think there is the problem?
I can't login to database because
SQL> connect /as sysdba
ERROR:
ORA-09817: Write to audit file failed.
Linux Error: 28: No space left on device
ORA-01075: you are currently logged on
So I must delete something.
Can I delete all core.* files to get space?
Why there will be so much core-dumpfiles?
In meantime I will looking for the target.xml file. 
core files existing under the directory/file system of cdump may be deleted. Unless you are Oracle support or a true GURU that knows how to make use of their information they are of no benefit.
...and yes, if the pw was changed and the xml file was not, OEM will report problems...
Edited by: KG_DBA on Oct 28, 2008 10:34 AM 
I am definitely not a guru so I will delete this dumpfiles ;).
But the main question is why the database produces so many dumpfiles since I have rebooted the machine...
Thanks for your help! 
I deleted now some dump-files and had a free space of 10G.
After that I created a new record: load: 332
The system wrote dumpfiles till no space was left.
-------
I can remember that the password would changed in grid control.
So I must change the password also in the targets.xml and set the value to false and write it in plaintext? 
The encrypted pw in targets.xml is the SYSTEM password for Oracle. If it has changed then yes you must make those changes in your XML file. Some places may differ at install time giving different pws for sys, system... 
I don't mean the database-password of the system or sys user.
Because with the password which I know I can log on as sys or system into the database.
The password which changed is for the communication between the agent of the clients and the grid control.
Now I stopped the database and all running processes.
Edited by: pairosilva on Oct 28, 2008 4:24 PM 
Lets be clear about this...
If you know the Oracle SYSTEM password was changed when your associate left, then that is the password that is required in targets.xml 
Ok but thats not the problem. I know all passwords and changed nothing.
In the targets.xml I type in the password but nothing changed.
In the meantime I found a big problem:
The emoms.properties is empty.
Yesterday I opened a SR at metalink maybe they can help me.
The support told me that my problem with the empty file is a bug but I don't get an answer to my question "why the oms produces 30Gbyte dumpfiles"
Kind Regards
René 
So the dump files you are talking about are the ones that go in directory/file structure CDUMP? If that is the case there should be a .trc file generated in UDUMP. Check your BDUMP alert log for generated ORA errors. You want to search here OR on the web for the details of the ORA error AND you need to review the .trc file that that error code points to.
As for the missing contents to emom.properties this will certainly affect OEM operations...worse case scenario...reinstall grid control but make certain all your host OS patches as defined in the installation documentation are present. 
No the files don't come from the database because the core_cump_dest is not the same path where the core dump will generated.
I shut down the database and the dumpfiles still already generated. So it makes no sense to check the bdump because it is definitely not a database problem.
When I start the OMS the dumps will generated.
Thank you for your help

Oracle 11g Install hangs when "creating clon database"

hi, i'm new with all this stuff, i'm trying to install oracle 11g on a windows 2003 standard edition and the installation hangs in this point:
http://img233.imageshack.us/img233/1458/installni7.jpg
it's almost an hour and it's stuck in 45% when trying to create clon database in "Creating and starting oracle instance".
I checked the log file but it says nothing weird, just that everything went well and a "restore" was made.
what can i do?
thanks!!! 
Check the alert.log file of that database being made.
usually under $ORACLE_BASE/admin/<SID>/bdump 
I have the same problem when installing,
I try several times ,and it always hangs on 45%
Edited by: yu2peng#gmail.com on 2009/2/18 上午 4:31
Edited by: yu2peng#gmail.com on 2009/2/18 上午 4:34 
If someone still needs an answer to this:
I had a similar problem with 10g. If you use certain special characters in the password for the database during its creation, some script will hang forever at 45% of the database creation. Just use a plain password without special characters. At the end of the database creation you can change all passwords. At that stage you can use special characters without any problem.
I don't know if this is the same in 11g, but it sounds very similar.
JD

No Images - Fresh install of V 2.18 in 32 bit Oracle Linux

Any Ideas?
I will add, this is Enterprise Linux AS Release 4 Update 8
Edited by: Alan3 on Mar 9, 2011 12:28 PM 
What do you mean "No Images" 
I go to http://<server>:8080
I get the login screen with no graphics... only place holders.
If I try to directly access one of the images
(e.g. http://<server>:8080/images/swan/FNDSSCORP.gif)
I get a 404 error. 
Please upload the contents of file mw_install_##########.log 
Sorry - I had already started a complete re-install.
After I posted, I found some issues with the listener setup. I can't prove it now, but I'm guessing that the installer never completed the database creation...
I cleaned off the previous install, created a new db and reinstalled all of it.
Everything seems to be fine now.
Thanks for your response, tho! 
can you send me an email letting me know the LISTENER problem that you encountered and then how you specifically resolved the LISTENER problem so that you could proceed and do the reinstall?
thanks
Michael
daltonmjd#yahoo.com

How to tell if database installation was successful?

Hi,
I think I've successfully installed Oracle Database 11.2 on Oracle Linux 5.7, because no non-ignorable critical errors occured. It didn't say "installation successful" either. I did a command-line (-silent) installation so maybe that's just how it worked. This is my first time attempting to install Oracle Database.
Anyway, how can I test to see if I've installed it? Any new commands I can check to see if they were installed? The documentation is way too overkill to follow. Does Oracle Database have any daemon processes that should be running? I do have an app folder that has a large hierarchy of files in it. 
898649 wrote:
Hi,
I think I've successfully installed Oracle Database 11.2 on Oracle Linux 5.7, because no non-ignorable critical errors occured. It didn't say "installation successful" either. I did a command-line (-silent) installation so maybe that's just how it worked. This is my first time attempting to install Oracle Database.
Anyway, how can I test to see if I've installed it? Any new commands I can check to see if they were installed? The documentation is way too overkill to follow. Does Oracle Database have any daemon processes that should be running? I do have an app folder that has a large hierarchy of files in it.If the installation started up, there should be a bunch of processes running, what they are is in the [url http://docs.oracle.com/cd/B28359_01/server.111/b28318/process.htm#CNCPT1246]docs. ps -ef (or -aux, whatever your OS uses)|grep pmon should tell you if it tried to come up. If no pmon, then be sure your ORACLE_HOME and ORACLE_SID are set, ORACLE_HOME/bin is on the path, you are logged on as oracle and try:
sqlplus / as sysdba
If it says "connected to an idle instance" then say startup. If it just gives a sql prompt, say select * from v$version;
There's a lot more, including the emctl stuff that lets you use browser access to see what is going on.
The docs are your friend. A needy, complicated, overweight, neurotic friend. But with benefits. 
Thank you sir. I think they did come up at first because my RAM filled up, so I restarted. Could you hint to me what the regular init command is to start them again? 
I did the following to set ORACLE_HOME and ORACLE_SID and PATH as you said:
$ export ORACLE_HOME="/home/oracledb/app/oracledb/product/11.2.0/dbhome_1"
$ export ORACLE_SID=501
$ export PATH=$PATH:$ORACLE_HOME/bin
Now when I try running sqlplus, I get an error.
$ sqlplus / as sysdba
sqlplus: error while loading shared libraries: libsqlplus.so: cannot open shared object file: No such file or directory
Edited by: 898649 on Nov 29, 2011 2:11 PM 
So, while I do see libsqlplus.so in $ORACLE_HOME/lib, it looks like it didn't get installed to my /usr/lib directory. Then again, I never ran the installer as root or anything, so it couldn't have. How does one deal with this?
The documentation said "During the database installation, when Oracle Universal Installer prompts you to run a script with root privileges, " but it never did this. I was running in -silent mode.
Edited by: 898649 on Nov 29, 2011 2:18 PM 
It's been way too long since I've dealt with silent installs and been on your platform, so someone else will have to help. There should be a root.sh script you should see what is in it, and instead of manually setting your environment look and see if there is an oraenv somewhere (like /etc or /opt) you can dot.
However, I do see MOS (the support site) notes like libaio.so.1: cannot open shared object file [ID 859523.1] which says this can come from not having all the required linux packages installed. There is a set of oracle-validated packages for rhel dists http://oss.oracle.com/el5/oracle-validated/ and see http://www.oracle.com/technetwork/server-storage/linux/articles/index.html
Note 1206898.1 mentions you have to set LD_LIBRARY_PATH for new client installs, and see Billy's excellent writeup: What is LD_LIBRARY_PATH env variable for? 
What options did you pass to the silent install ? Can you pl post the last 100 lines of the install log ?
http://docs.oracle.com/cd/E11882_01/install.112/e24321/app_ts.htm#i1090466
Did you verify that you met all of the pre-requisites ?
http://docs.oracle.com/cd/E11882_01/install.112/e24321/pre_install.htm
HTH
Srini 
Hi,
1. Check installation log file
2. Check result for below query and post it here if possible
SQL> select comp_name, version, status from dba_registry;
Regard
Helios 
notfed wrote:
Hi,
I think I've successfully installed Oracle Database 11.2 on Oracle Linux 5.7, because no non-ignorable critical errors occured. It didn't say "installation successful" either. I did a command-line (-silent) installation so maybe that's just how it worked. This is my first time attempting to install Oracle Database.
Anyway, how can I test to see if I've installed it? Any new commands I can check to see if they were installed? The documentation is way too overkill to follow. Does Oracle Database have any daemon processes that should be running? I do have an app folder that has a large hierarchy of files in it.You didn't say if your installation included the creation of an actual database or if it was just the rdbms software.
In any event, why not assume it was good and proceed accordingly? If it wasn't good, you'll find out soon enough, and then can work the specific issue. 
notfed wrote:
So, while I do see libsqlplus.so in $ORACLE_HOME/lib, it looks like it didn't get installed to my /usr/lib directory. Then again, I never ran the installer as root or anything, so it couldn't have. How does one deal with this?
The documentation said "During the database installation, when Oracle Universal Installer prompts you to run a script with root privileges, " but it never did this. I was running in -silent mode.
Edited by: 898649 on Nov 29, 2011 2:18 PMIf you want to run the root.sh, you just have xecute $ORACLE_HOME/root.sh as root. 
I decided to start over and try again. New post: What's wrong with my Oracle Database Installation?

11gR2 em install / config question.

We just loaded 11gR2 onto a virtual server running Win 2008 Server R2. During the installation of the database,
I received 5 or 6 missing file messages and I told the installer to ignore and continue. One of those files was
dbhome_1\...\em.jar.
Will this preclude me from getting em up? Should I deinstall/reinstall the 11gR2?
When I ran emctl start dbconsole I received this message:
can't locate CompEMdbconsole.pm in #inc <#inc contains ..(list of files).......emctlCommon.pm line 598
From this question you'll probably surmise that I'm relatively inexperienced.
Suggested reading info is always appreciated.
Thanks,
Ron
Edited by: RonW on Oct 9, 2012 6:27 AM 
Well, looks like your OEM installation was not succesfully.
It would be really import to get this 5/6 missing messages.
I suggest you to run dbca and remove/reinstall OEM.
Regards, 
Thanks for the reply. I used dbca to create a database which I had thought was done
during the original install. It ran forever and eventually gave me an error message
'Error Instantiating OC4J files'. When I looked in the emconfig.log file, there were
several missing xml files in E:\app\Administrator\product\11.2.0\dbhome_1\oc4j\j2ee\
OC4J_DBConsole\config\
jms.xml
rm1.xml
http-web-site.xml
server.xml
In fact, the 'config' folder was empty. Going back to my original question regarding
those missing files, why would they be missing? Is the oem a separate install
process or is it done when the 11g database is installed?
Thanks, Ron 
Hi,
It is optional in the installation process. The missing files may be issue from your installation.
But anytime you can use the dbca and install OEM. You should pay attention, i think on the installation process the database would have to be offline, check it out.
Regards,

Categories

Resources