Patching 10gr2 RAC, Oracle inventory on a CFS

I have 4-node HP-UX VCS5 Oracle10g RAC setup running in production environment. First time i am trying to apply latest patch bundle. I am trying to apply 10.2.0.4 patch set.
I am getting the bellow error when i am applying the patch.
"Central Invventory can't be shared, this is not supported"
The ORACLE_BASE is on a cluster file system.
To ignore this error and to proceed with patch installation, what steps we need to take ?
Thank you.

Hi,
Looks like you have run into an always documented restriction which has now been enforced with the OUI in the 10.2.0.4 patchset.
The documentation which mentioned that Shared Central Inventories are not recommended / supported is available at :
http://download.oracle.com/docs/cd/B19306_01/em.102/b16227/oui6_cluster_environment.htm#BABGJAAF
<Quote>
The Oracle Inventory directory (oraInventory) stores an inventory of all software installed on the system. It is required by, and shared by, all Oracle software installations on a single system. The first time you install Oracle software on a system, Oracle Universal Installer prompts you to specify the path to this directory. If you are installing the software on a local file system, then Oracle recommends that you choose the following path:
oracle_base/oraInventory
If the Oracle base directory is on a cluster file system, or on an NFS file system on a NAS device, then you must specify a path for the Oracle Inventory directory on a local file system. The Oracle base directory must be on a local file system to enable all of the nodes to have separate inventories.
Oracle Universal Installer creates the directory that you specify, and sets the correct owner, group, and permissions for it. You do not need to create it.
<un-quote>
What has changed with the 10.2.0.4 Installer is that Oracle now enforces this restriction. What I believe is that without this being enforced , there was a possibility of running into Central Inventory corruption issues when Clustered and Non-Clustered installs were being done on the same node.
I also found this discussion on one of the Metalink threads which might be useful.
Please contact Oracle Support prior to attempting these steps though.
Quoting Oracle:
It is not supported to have central inventory on a shared location. There was a bug.5009954 "OUI SHOULD ENFORCE SEPARATE CENTRAL INVENTORY LOCATIONS IN RAC" filled on this issue. This bug is fixed in 11.2. It is actually an OUI issue which doesn't perform a check to make sure that central inventory location is not on shared device.
In order to fix the inventory, you can change the OH/oraInst.loc in the CRS home to point to a non-shared directory.
Run <crs>/oui/bin/attachhome.sh -silent -local -cfs CRS=true -invPtrLoc <crs>/oraInst.loc on each of the nodes. Finally copy the oraInst.loc to the /etc/oraInst.loc.
Vishwa

Similar Messages

Solaris 10 + Oracle 10gR2 RAC question

Hello everyone
Has anyone come across the case where the CRS services of Oracle cause
the public interface to get turned off and then restored at random
time intervals? To elaborate, we have a 2 node cluster database.
Solaris 10, Oracle 10gR2 RAC with patch 10.2.0.3 applied. No SUN
clustering is involved. When the cluster software is down (nodeapps,
asm, database instances all down) /var/adm/messages show nothing. When
we start nodeapps on the 2 nodes(thus initiating some form of
communication between the nodes), at random time intervals we get
"interface ce0 turned off and interface ce0 restored" in /var/adm/
messages. When we check the status of the RAC, we see that one node's
vip has been assigned to the other. This on/off behaviour of the NIC
can be eliminated only if we continuously PING it from a another
client in the network.
As a matter of fact, the RAC and the RDBMS work perfectly when we keep
pinging the 2 nodes from an other client on the network. We even
managed to run a long batch job, distributed on cluster managed
services on the 2 instances, and it completed after 9 hours without
any problems.
Does anyone have a hint on this behaviour? Is there some sort of
timeout for the network cards? Some power saving features? Googling
around I came across the new Containers feature available on Solaris
10. Is there a way that I can verify that either RAC or the RDBMS is
running in "container" mode ( since the solaris and Oracle
installation was not performed by me)? Any other ideas?
Thank you for reading

Im an Oracle guy - not the SA type -
But on ours - the SA configured this cluster incorrectly. We use veritas. instead of making ipmp groups for the interfaces - he built the cluster according to veritas docs. That is - he has two publics - on difference interfaces and different privates on different interfaces. oracle can only two interfaces - no matter if its a ipmp group or a device name. one is used for private- the other is used for public - So sure the veritas cluster filesystems will survive - but the Oracle Cluster will not - nodes will reboot -
Is your system set up incorrectly as i described above? if it is - a quick test would be - turn down the other interfaces - and only leave the two interfaces you mention above up that you configured for Oracle CRS
This other sharp sa was able to go through the arp table and see duplicate IPs - and the routing was attempted via an interface that oracle doesnt see. You can not define two different interfaces public - and two different interfaces for private -

Upgrade Oracle 10gR2 RAC to 11gR2 RAC

Hi Gurus,
I got a requirement to upgrade my existing environment from Oracle 10gR2 RAC to 11gR2 RAC. My environment details are as follows
Operating system RHEL 5.5
Cluster version 10.2.0.4
ASM version 10.2.0.4
Database version 10.2.0.4
Storage HP EVA (Multipathing)
Could any one please specify the links or documents to complete this task!
Thanks & Regards,
Venkat

HI!
In my practise we have hardware to move 10g -> 11g.
We upgrade our DB's -> install Grid 11g on new server, move DB 10g, Upgrade 10g -> 11g.
You should work on this topic's:
Oracle® Database Upgrade Guide 11g Release 2 (11.2)
NEW FEATURES 11gr2
Complete Checklist for Manual Upgrades to 11gR2
Oracle Recommended Patches
Oracle Clusterware - ASM - Database Version Compatibility
How to Download and Run Oracle's Database Pre-Upgrade Utility
Best Practices to Minimize Downtime During Upgrade
Things to Consider Before Upgrading to 11.2.0.2 Grid Infrastructure
Different Upgrade Methods For Upgrading Your Database
All in you hands!!!

Oracle 10gR2 RAC - ASM question

Hi
I have a question regarding the ASM storage. Let says I have a system here running Oracle 10gR2 RAC and would like to add a new/extend the current DATA disk group with more disk space. How do I do that? will it affect the existing data stored inside there?

So to add a little more to the discussion. Let's say your storage administrator presents you a LUN and is nice enough to add a partition of say 7G. (/dev/sdo1).
Now you need to take /dev/sdo1 stamp it and alter your storage group.
For illustration purposes I shall use rac1 and rac2 as my dual instance RAC and add to the asm group ARCH.
As root on rac1
/etc/init.d/oracleasm createdisk ARCH2 /dev/sdo1
then run
/etc/init.d/oracleasm listdisks
to make sure ARCH2 shows up.
On rac2 you run
/etc/init.d/oracleasm listdisks
You don't see ARCH2 so then run
/etc/init.d/oracleasm scandisks
then
/etc/init.d/oracleasm listdisks
Now you should see ARCH2
Ok the asm stamps are in sync now.
Back to rac1
su - oracle
set ORACLE_SID to asm instance and use sqlplus
sqlplus / as sysasm
If you query V$ASM_DISK you will see your disk with a header_status of PROVISIONED
that's good ...
NOw while still in sqlplus
Let's bump up the asm_power_limit so rebalancing runs faster
alter system set asm_power_limit=5 scope=both ;
If your asm instance are sharing the same spfile you only need do this on one instance; otherwise run the command both on all asm instances.
Lastly
ALTER DISKGROUP ARCH ADD DISK 'ORCL:ARCH2' ;
Now you can query V$ASM_OPERATION and watch ASM do it's magic of rebalancing.
That's it. All done while the DB is up and running.
How does that work for you?
-JR jr

An issue for 10gR2 RAC patching to 10.2.0.3

Hi, all:
I am patching RAC (two nodes) database from 10.2.0.1 to 10.2.0.3
one thing I want to verify:
for preinstalltaton, I stop all processes in RAC, one of them is to stop applications---
srvctl stop nodeapps -n <node>
(that is required by manul)
after pacthing, and then after DBUA to update instance to 10.2.0.3,
do I need to do:
srvctl start nodeapps -n <node>
I didn't see starting that after pacthing in patching manual provided by Oracle metalink?
also, welcome any friends' experience for this kind of patching/upgrading?
thank you very much!!!

If you took the "Oracle Databsae 10g:RAC for Administrators" Oracle Curriculum course, at chapter 12 there is a nice lab where a practical database patching operation is carried out.
It is a practical step by step patch on rac environments.
~ Madrid.

10gR2 RAC On Solaris9

is 10gR2 RAC On Solaris9 is certified by oracle? If i install it on Solaris9, what problems i could face? & what would be the advantage of installing this on Solaris10?

You can check the matrix certification without metalink access :
http://www.oracle.com/technology/support/metalink/index.html
A certified config means first that works (tested by Oracle Corp.), then Oracle support team is allowed to answer to your issues, lastly and not at least, Oracle may build out patch/patchset if required.
Nicolas.

'unable to find product oracle.swd.jre[1.3.1,1.6] in Oracle Inventory'

Hi All,
i am running the runInstaller for patch set 9.2.0.8 on AIX 5L Based Systems (64-Bit)
after selecting the products.xml file and selecting the oracle home to update and then cllick Install
i get the error 'unable to find product oracle.swd.jre[1.3.1,1.6] in Oracle Inventory'
is it the issue of the version of oracle.swd.jre? can we skip the installation of this componenet?
Kindly suggest
Thanks
SK

Hi,
Please see if these documents help.
Note: 397745.1 - Error when installing Patch 2376472: "Unable to find product oracle.swd.jre "
Note: 315842.1 - 'unable to find product oracle.swd.jre' Error During Agent Upgrade from 10.1.0.3 to 10.1.0.4
Regards,
Hussein

10gR2 RAC on Solaris 10

Hi,
I am looking for "step by step install & configure" guide to setup 10gr2 RAC on Solaris 10.
I got the following Doc "ID: Note:317257.1" from metalink. Looks like I need to configure Solaris 10 containers and Solaris Zones for 10g RAC.
Is there another way to setup 10gr2 RAC on Solaris 10 without using the Solaris 10 containers just like (Solaris 9 or plain OS)?
Any help would be appreciated.
R-

Thanks for all your inputs. I finally got 10g RAC installed and running.
I learned that there is no way to install Oracle software and configure ASM during installation. I had to install “Software Only” then configure ASM at the later time, then that will work.
Good luck to you all.
:-)R

Can we have separate 10gR2 RAC CRS and 11gR2 RAC CRS homes on same Unix s

We'd like to test the upgrade of our 10gR2 RAC databases to 11gR2 RAC.
We have six databases on the same server and want to install a separate 11g RAC stack (11g CRS, 11g ASM, 11g RDBMS) and upgrade each 10g database from 10g RAC stack on same server (10g CRS, 10g ASM, 10g RDBMS) . Is it ok to have 10g CRS home and 11g CRS home same server? If not, how best to upgrade without buying more hardware?
Thanks in advance,
Lou

In more detail, the main issue/question is whether there can be 2 Oracle CRS homes on the same server i.e. separate 10gR2 and 11gR2 CRS homes residing on the same server for the purposes of (gradually) upgrading all Oracle RDBMS homes from 10g RAC to 11g RAC.
Can there be? What would be the implications for the OCR and voting disk? In our case 10g CRS has OCR and Voting Disk on raw devices. In 11gR2, these would need to be migrated/upgraded/stored in ASM, right?
We have 6 10g rdbms homes in addition to the one 10g CRS home and one 10g ASM home. To upgrade can we
- Install 11gR2 CRS in a separate home (this would now include ASM as well)
- Install 11gR2 RDBMS home
- Upgrade one of the 6 10g databases while leaving the others untouched as 10g databases in the separate 10g CRS, 10g ASM stack?
Any ideas, pointers greatly appreciated.
Thanks,
Lou

Oracle inventory location Help

I am facing some issues with oracle inventory locatoin on a 2 node rac setup. The ORACLE_BASE in my .bash_profile is pointing to /u01/ mount point - which is a NFS mounted volume. clusterware installation went sucesful and now when i am trying to install 11g - it says, oracle inventory can not be on a shared volume and quits the installation.
My oracle home directory is \home\oracle\.
system has three mounted directories: /u01/ - /u02/ - /u03/.
/u01/app/oracle/crs - clusterware
/u01/app/oracle/product/11g/ - plan on installaing oracle 11g
/etc/oraInst.log - has the inventory location.
can somebody please suggest - what options do i have here? any input will be greatly appreciated.
many thanks,
shilpa

Oracle Inventory can not store on Share location.
http://kevinclosson.wordpress.com/2007/07/19/oracle11g-oracle-inventory-on-shared-storage-dont-bother-trying-to-install-11g-rac-that-way/

Install 10gR2 RAC w/ iscsi chokes "filesystem too small for journal"

Following Jeff Hunter guide to install RAC on 10gR2 on Oracle Linux All has gone well to the point of:
[root@lin1 ~]# mkfs.ocfs2 -b 4k -C 32K -N 4 -L oracrsfiles /dev/sde1
mkfs.ocfs2 1.2.4
Filesystem too small for a journal
[root@lin1 firefox]# df -m
Filesystem 1M-blocks Used Available Use% Mounted on
/dev/mapper/VolGroup00-LogVol00 66272 1965 60941 4% /
/dev/cciss/c0d0p1 99 16 79 17% /boot
none 1 268 0 1268 0% /dev/shm
[root@lin1 ~]# uname -a
Linux lin1 2.6.9-55.0.0.0.2.ELsmp #1 SMP Wed May 2 14:59:56 PDT 2007 i686 i686 i386 GNU/LinGNU/Linux
Tried man mkfs.ocfs2 with the -J size=(various sizes in KB up to 100MB) (4MB is the min) but nothing takes. Certainly approx. 6GB free space is sufficient for journaling?? Any input very much appreciated.

Question related to DB patching of RAC from 10201 to 10205

Hello,
I have the task of patching a RAC environment from 10201 to 10205 on linux x86_64, as per the README for the patch (#8202632) I have to first apply the patch to
1.Clusterware.
2.Oracle database homes.
I am a bit new to patching RAC databases , just wanted a confirmation that the database (even if the ORACLE_HOME are not shared and are local to each node) patch is not a rolling patch.*.The README says
Shut down all Oracle RAC instances on the nodes, which run from the Oracle home on which you are going to apply the patch set. To shut down all Oracle RAC instances for a database, enter the following command where db_name is the name of the database:
*$ srvctl stop database -d db_name*
The above command will shutdown the databases on all the nodes, thus it is not a rolling upgrade and the database has to be down, so the end users have to informed prior to patching that there is a down time involved.
Any confirmation from folks who have already done this will be much appreciated.
Thanks,
C.

Hi,
For clusterware it can be rolling patching.
- apply the clusterware 10.2.0.5
- follow the instruction at the end of installation
for the database patching it can't be rolling upgrade means database must be down
CHeers

Corruption of Oracle Inventory by installing ADF runtime libraries?

Hi,
We are running OAS 10.1.2.2.0 on RedHat Enterprise Linux and I am deploying an ADF application from JDev 10.1.3.39.84.
I downloaded the standalone installer for the adf runtime libraries installer from here:
http://download.oracle.com/otn/java/jdeveloper/10132/extensions/adfinstaller.zip
and installed it to our development server. I then deployed the ADF application and it seems to work just fine.
Now I have requested that these libraries be installed to the production server but the Sys Admin has informed me that:
"Your installation appears to have corrupted the 'Oracle Inventory' on the development server. This can affect upgrades, patching, etc."
I have been asked to investigate if this is a known issue and what possible resolutions there might be for preserving/restoring the Oracle Inventory to enable continued maintenance of the servers.
I have searched the forums and Metalink without success, so I am posting this query to see if anyone else can help me point me in the right direction.
Thanks for any assistance.

Hi,
please file a service request with customer support. This way you have an analyst working with you to track down the issue. Because ADF libraries are copying jar files only and this is nothing you can re-do after wards, I would be surprised if it has anything to do with the Oracle inventory. However still it would be good to have customer support looking at this
Frank

10gR2 RAC installation: OUI forgets to create directory 'crs/auth' (BUG)

Hi,
We've got a new installed 10gR2 RAC cluster (2 nodes) on AIX. But there were several errors during startup of CRS (Cluster Ready Services).
The error log shows, that it couldn't write a special file (I guess a lock file).
So I just created the missing directory 'crs/auth' and anything works fine now.
This post is just for those who are looking for a solution for their non-working crs daemon. No need to answer.
But I suppose Oracle should log it as an installation bug, because the Universal Installer does not create this directory and so the CRS can not work.
Best regards
Sven

it looks like i hit a bug in crs 11.1.0.6 -> bug 6782573
becuase i can see in my /var/log/messages
Jul 30 00:53:07 rac1 logger: Oracle CRSD respawn disabled. Unsupported change in hostname.
Jul 30 00:53:08 rac1 logger: Oracle init script ceding reboot to sibling 10986.
Jul 30 00:53:08 rac1 logger: Oracle EVMD respawn disabled. Unsupported change in hostname.
Jul 30 00:53:12 rac1 logger: Cluster Ready Services completed waiting on dependencies.
Jul 30 00:53:13 rac1 logger: Cluster Ready Services completed waiting on dependencies.
Jul 30 00:53:14 rac1 logger: Running CRSD with TZ =
Jul 30 00:55:08 rac1 logger: Oprocd received graceful shutdown request. Shutting down.
Jul 30 01:58:07 rac1 logger: Oracle CSS family monitor shutting down gracefully.Jul 30 02:06:38 rac1 logger: Oracle EVMD respawn disabled. Unsupported change in hostname.

Guidelines for SGA and PGA tuning for 10gR2 RAC and ASM ?

I am looking for tuning information on SGA and PGA tuning for Oracle 10g (10.2) RAC and ASM for a 4TB OLTP and DSS mixed environment on Solaris 10 platform.

We are running Solaris 10 SPARC 64 bit with Oracle 10gR2 RAC Enterprise Edition and ASM on Sun servers with 32GB of RAM for memory for a 4TB OLTP database.
It is in design phase so I do not have an existing AWR or Statspack report yet. Is there a best practices guideline on how to size parameters for the SGA (ie: shared_pool_size, etc) and PGA for this environment from Oracle?

Patching 10gr2 RAC, Oracle inventory on a CFS

Similar Messages

Maybe you are looking for