Oracle® Database

4.1.1 Disk Space Requirement Issue During Oracle Grid Infrastructure Installation

When installing Oracle Grid Infrastructure, the disk space requirement displayed on the Oracle Universal Installer Summary page is less than the actual disk space required.

Workaround:

Refer to Oracle Grid Infrastructure Installation Guide for information about disk space.

This issue is tracked with Oracle bug 9241366.

4.2.1 Microsoft Visual C++ Redistributable Packages

The Microsoft Visual C++ 2005 SP1 Redistributable Package is automatically installed with the Oracle Database 11g Release 2 (11.2) software. Do not remove this package from the Add or Remove Programs interface in the Control Panel.

4.2.2 Upgrading Windows 32-Bit to Windows 64-Bit Using Oracle OLAP

There are specific instructions for migrating 32-bit Oracle Online Analytical Processing (OLAP) option databases to 64-bit. These instructions are documented in Note 352306.1 on My Oracle Support at

https://support.oracle.com/

4.2.3 Issues Associated with Oracle SID Containing Underscore Character

Oracle Database 11.2 supports Oracle SID having an underscore but it is not compatible with the earlier client releases with a bequeath connection. The issue affects any Oracle Database Client 11.1.0.7 or earlier client, when connecting to the database over a bequeath connection. This includes all OCI clients and Oracle Enterprise Manager agents that monitor the database.

Workaround:

Use only alphanumeric characters including hyphens for Oracle SID or connect to the database using TCP connection.

This issue is tracked with Oracle bug 7611024.

4.2.4 Enabling Oracle Automatic Storage Management Dynamic Volume Manager Volumes

Oracle Automatic Storage Management Dynamic Volume Manager volumes are typically enabled automatically. Disk groups that are listed in the init.ora file for Oracle Automatic Storage Management (Oracle ASM) have all volumes automatically enabled during disk mount.

For volumes in disk groups that are mounted manually, you must also manually enable the volumes in that disk group before using them.

In Oracle Restart, to load Oracle Automatic Storage Management Cluster File System drivers during system initialization as an Administrator, you must manually reload the three drivers associated with Oracle ACFS: Oracle OKS, Oracle ADVM and Oracle ACFS.

This issue is tracked with Oracle bug 7700977.

4.2.5 Automatic Storage Management Service Is Disabled After System Restart

After an installation, the Automatic Storage Management Startup Type is set to Manual. Hence, Automatic Storage Management does not start when you restart the system.

Workaround:

In Services Control Manager, change the Startup Type to Automatic.

4.2.6 OCI Instant Client Version Update Issue

A new Oracle Database release does not update OCI Instant Client version.

Solution:

Use the OCIClientVersion feature to determine the client version instead of checking the dll version number.

This issue is tracked with Oracle bug 8372062.

4.2.7 ORA-00600 Error During Database Upgrade

When you upgrade from Oracle Database Release 9.2 to release 10.2, and then upgrade to release 11.2, the database upgrade fails with an ORA-00600 error.

This issue is tracked with Oracle bug 8668823.

4.2.8 Oracle Clusterware Upgrade Error

When you use Oracle Universal Installer to upgrade Oracle Clusterware 11g Release 1 (11.1.0.6), or Oracle Clusterware 11g Release 1 (11.1.0.7) to Oracle Clusterware 11g Release 2 (11.2), you may receive the following error:

[INS-40406] There is no previous version of grid software detected on the system.

Workaround:

If you are performing the upgrade on a remote node, a node other than the one on which Oracle Clusterware 11g Release 1 (11.1.0.6) was installed, then the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc may not have been set on that node.

Ensure that you manually set the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc to the location of the Oracle Inventory directory of the node on which you are performing the upgrade installation.

This issue is tracked with Oracle bugs 9210861 and 6527941.

4.2.9 EMVLOGGER.EXE Process Does Not Restart Automatically

If the EMVLOGGER.EXE process is terminated at the operating system level, it fails to restart automatically because of an existing log file.

Workaround:

Remove the emvlogger log file at Grid_home\evm\log\*_evmlogger.log, where Grid_home is the location of your Oracle Grid Infrastructure installation. The evmlogger.exe process should automatically restart after approximately one minute.

This issue is tracked with Oracle bug 9260650.

4.2.10 Windows Service Issue for Oracle Database

Manually shutting down an Oracle Database or stopping an Oracle ASM instance using CRSCTL utility or SQL*Plus does not stop its associated Windows service.

Workaround:

Using the Services management console, manually stop the Windows service associated with the Oracle instance. You can access the Services management console through the Control Panel, or you can stop the Oracle Database service, or Oracle ASM service directly using the following commands:

net stop OracleServiceSID
net stop oracleASMServiceSID

This issue is tracked with Oracle bug 9178097.

4.2.11 Remote Agents Must Be Manually Reloaded After Cluster Ready Services Upgrade and Oracle ASM Upgrade

When attempting to upgrade Cluster Ready Services or Oracle ASM to 11.2 release using the Automatic Storage Management Configuration Assistant, the upgrade succeeds, but the configuration changes made to the Agents on a remote host fail to reload. As a result, Oracle Enterprise Manager Grid Control and Database Control cannot monitor the Oracle ASM and Cluster Ready Services targets.

Workaround:

Manually reload the remote agents after the upgrade using the following command:

emctl reload agent

This issue is tracked with Oracle bug 9398380.

4.2.12 Oracle MTS Recovery Service Issue When Installing Oracle RAC

In an Oracle RAC installation, Oracle MTS Recovery Service on the remote nodes does not have the correct host name registered for string variable HOST under registry key HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0. This variable contains the name of the host from where the installation was performed instead of the name of the local host. This registry variable is used by COM+ applications to pass information to Oracle databases so that they can connect back to the Windows computer if the database must resolve in-doubt COM+ transactions. The incorrect host name that is registered can cause transaction recovery issues.

Workaround:

To fix this, use Registry Editor (regedit) on the remote Windows system and look under the registry key:

HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0

The string-type values for the following key must be set correctly:

Name:            *Host*
Type:            REG_SZ
Data: << the name of the host-machine e.g. myhost@example.com <mailto:myhost@example.com>>> 

This key must contain the host name of the system on which the registry key exists.

This issue is tracked with Oracle bug 9483665.

4.2.13 Root Script Status Not Displayed with Oracle Grid Infrastructure

When running rootcrs.pl or roothas.pl to configure, upgrade, or deconfigure Oracle Grid Infrastructure, the script does not display the success or failure status in the command window.

Workaround:

None. Check the log file generated at the following location to know the status of the script:

Grid_home/cfgtoollogs/rootcrs_host_name.log 

This issue is tracked with Oracle bug 9473262.

4.2.14 Issue Encountered When Special Characters Are Used in Network Interface Names

While installing or upgrading Oracle Grid Infrastructure for a cluster, you might encounter an issue if special characters, including non-ASCII characters and single quotation marks, are used in network connection names. The following message is recorded in the Oracle Universal Installer log:

ATTENTION: Bareword found where operator expected at (eval 20) line 1,
near "'"node-vip/netmask/excutabled'acc"
ATTENTION:     (Missing operator before acc?)
ATTENTION: Parameter ORACLE_HOME not defined at
C:\app\11.2.0\grid\crs\install/crsconfig_lib.pm line 8722. 

Workaround:

Do not use special characters in network connection names.

This issue is tracked with Oracle bug 9492881.

4.2.15 Deinstallation Tool Might Not Remove ORACLECLUSTERVOLUMESERVICE

When using the Deinstallation Tool to remove a Clusterware home, you might encounter the issue that removing service, OracleClusterVolumeService, failed.

Workaround:

Manually remove the service, OracleClusterVolumeService.

This issue is tracked with Oracle bug 9405729.

4.2.16 Deinstallation Tool Fails to Remove Registry Entries

When deinstalling a single instance Oracle Database, Oracle Database Client, or Oracle RAC (only 64-bit Windows servers), some registry entries are not removed.

Workaround:

• Manually remove the following registry entries:

  HKLM\Software\Microsoft\.NETFramework\v2.0.50727\AssemblyFoldersEX\ODP.NET
  HKLM\Software\Microsoft\VisualStudio\8.0\Help\VisibleFilters\OracleDataProviderTechFilter
  

  This issue is tracked with Oracle bug 9501499.

• Manually remove the registry entry, Oracle in ORACLE_HOME_NAME:

  HKLM\Software\ODBC\ODBCINST.INI\ODBC Drivers\Oracle in Oracle_HOME_NAME
  

  This issue is tracked with Oracle bug 9478171.

• Manually remove the registry entry, Oracle Services for MTS:

  HKLM\System\CurrentControlSet\Services\EventLog\Application\Oracle Services for MTS
   
  

  This issue is tracked with Oracle bug 9524210.

• Manually remove the registry entry, OracleDBConsoleSID:

  HKLM\System\CurrentControlSet\Services\EventLog\Application
   
  

  This issue is tracked with Oracle bug 9523875.

• Manually remove the registry entries, Oracle.SID and Oracle.VSSWriter.SID:

  HKLM\System\CurrentControlSet\Services\EventLog\Application
   
  

  This issue is tracked with Oracle bug 9523866.

• Manually remove the following registry entries:

  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Oracle11
  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\OraFenceService 
  

  This issue is tracked with Oracle bug 9478187.

4.2.17 Deinstallation Issues with ODP.NET

While removing a single instance Oracle Database, Oracle Database Client, or Oracle RAC on 64-bit Windows, you might see some unnecessary characters or lines printed in winprod_cleanup.out on a non-English operating system. If the operating system is English, then you might see the following message:

This action is only valid for products that are currently installed.

Workaround:

Ignore these unnecessary characters and the message.

This issue is tracked with Oracle bug 9440539.

4.2.18 Deinstallation Tool Does Not Remove Oracle Home

When using the Deinstallation Tool to remove an Oracle home that has Oracle Enterprise Manager Database Control configured and that was installed on Oracle ACFS, the ORACLE_BASE\ORACLE_HOME\host_sid\sysman directory and its sub directories cannot be deleted, preventing the Oracle home from being removed.

Workaround:

Before performing these operations, run the following command as the Administrator user:

cacls ORACLE_BASE\ORACLE_HOME\HOSTNAME_DBNAME\sysman\emd\upload\hsperfdata_SYSTEM /E /P %USERDOMAIN%\%USERNAME%:F 

This issue is tracked with Oracle bug 9444816.

4.2.19 Deinstallation Tool Does Not Remove invDetails.properties File

The Deinstallation Tool fails to remove the invDetails.properties file when deinstalling Oracle RAC home.

Workaround:

Manually remove the dbhome\inventory directory.

This issue is tracked with Oracle bug 9543109.

4.2.20 Deinstallation Tool Does Not Remove ocfs.sys File

The Deinstallation Tool fails to remove the ocfs.sys file when deinstalling Oracle RAC.

Workaround:

Manually remove the ocfs.sys file:

system_drive\Windows\System32\drivers\ocfs.sys

This issue is tracked with Oracle bug 9498036.

4.2.21 Deinstallation Tool Does Not Detect Oracle RAC Home

If you are deinstalling the 11.2.0.1.0 Oracle RAC software installed on Oracle ACFS mounted drive using the Deinstallation Tool, you might see the following message during the Oracle Configuration Manager (OCM) deconfiguration:

Oracle Home does not exists, but CCR is configured

Workaround:

You can ignore this message, the OCM gets deconfigured successfully.

This issue is tracked with Oracle bug 9545478.

4.2.22 Using Deinstallation Tool for Oracle RAC Returns Errors

The Deinstallation Tool cannot delete files that are still in use because one or more processes or services were not stopped successfully. The deinstall output shows error messages such as:

Failed to delete the file file_name

Workaround:

Remove the Oracle home manually after the server is restarted.

This issue is tracked with Oracle bug 9446472.

4.2.23 Time Issue with Deinstallation Tool to Remove Oracle RAC Home

When using the Deinstallation Tool to remove an Oracle RAC home that was installed on Oracle ACFS, the tool takes a longer amount of time than expected to finish.

Workaround:

None. Wait an additional 20-30 minutes for the deinstallation tool to finish.

This issue is tracked with Oracle bug 9478372.

4.2.24 Deinstallation Tool Fails to Remove Oracle Grid Infrastructure for a Cluster in Silent Mode

When using the Deinstallation Tool to remove Oracle Grid Infrastructure for a cluster in silent mode, the tool fails with the following error message:

ERROR:  As the privileged user, Execute \deinstall\crsdc.bat on every
node  in this cluster except for this node, and then execute
\deinstall\rootdeinstall.bat on this node to remove the configuration of the
Oracle Clusterware and to completely remove the Oracle software 

Workaround:

Follow these steps when the deinstallation fails:

1. Run crsdc.bat on all the nodes as instructed by the Deinstallation Tool.

2. Run rootdeinstall.bat from the deinstall directory on node 1.

This issue is tracked with Oracle bug 9540713.

4.2.25 Install 32-Bit and 64-Bit Oracle Software in Separate Oracle Base Directories

If you install 32-bit Oracle software and 64-bit Oracle software on the same server, then you should use different Oracle Base directories for the 32-bit Oracle software and the 64-bit Oracle software.

This issue is tracked with Oracle bug 9502109.

4.2.26 User Issue with ORA_DBA Group on Remote Nodes

When installing an Oracle RAC on a cluster using a local user (non-domain user) account, the user account is not added to the ORA_DBA group on remote nodes. This prevents the local user from connecting to the database as SYSDBA using operating system authentication.

Workaround:

Manually add this local user to the ORA_DBA group on all the remote nodes or install as a domain user.

This issue is tracked with Oracle bug 9524516.

4.2.27 Oracle Clusterware Resource Status Issue

In an Oracle RAC cluster with 2 or more nodes, if you restart one node, and then, while the node is being restarted, use CRSCTl on another node to query the status of the Oracle Clusterware resources, then CRSCTL either displays no information, or returns the error message:

CRS-0245:USER DOESN'T HAVE ENOUGH PRIVILEGE TO QUERY CRS RESOURCES

Workaround:

Before restarting a node, issue the following command on that node:

crsctl stop crs -f

This issue is tracked with Oracle bug 9369059

4.2.28 Oracle Single Client Access Name (SCAN) Issue with Oracle Clusterware

In an Oracle Clusterware environment where DHCP is configured and GNS is enabled, if the non-qualified SCAN is resolved by DNS, it can cause Enterprise Manager Configuration Assistant or Database Configuration Assistant to fail with the following error message:

SEVERE: Scan Listener is not up or database service is not registered with it.
Start the Scan Listener and register database service and run EM Configuration Assistant again

Workaround:

When using GNS, ensure that the SCAN, qualified by either the default domain on the network, or qualified by any domain in the cluster nodes' domain search order, is not resolvable through DNS for the network.

This issue is tracked with Oracle bug 9471408.

4.2.29 Oracle Notification Service Fails to Start During Oracle Grid Infrastructure Upgrade

During an upgrade of the Grid home to release 11.2, the PRCR-1064 error is sometimes encountered for the ora.ons resource on the last node. This error appears in the rootcrs_nodename.log file.

Workaround:

When Oracle Grid Infrastructure upgrade fails with this error, perform the following steps:

1. Open the Windows Task Manager and find the two ons.exe processes.

2. Stop both the ons.exe processes.

3. Run the following commands:

   srvctl start nodeapps
   srvctl start gns
   srvctl start scan
   srvctl start scan_listener
   

4. Skip the Oracle Grid Infrastructure configuration from the installer and complete the upgrade.

This issue is tracked with Oracle bug 9491140.

4.2.30 Upgrade of Oracle Clusterware Release 10.2.0.5 to Oracle Clusterware 11.2.0.1 Fails

During an Oracle Grid Infrastructure for a cluster installation, upgrade of Oracle Clusterware 10.2.0.5 to Oracle Clusterware 11.2.0.1 fails. The installer displays the following message:

Grid Infrastructure Configuration failed.

Solution:

Upgrading directly from Oracle Clusterware 10.2.0.5 to Oracle Clusterware 11.2.0.1 is not supported. Either complete the upgrade by running an Oracle Grid Infrastructure for a cluster 11.2.0.2 installation, or contact support for a one-off patch for the Oracle Clusterware 11.2.0.1 software.

This issue is tracked with Oracle bug 9776578.

4.3.1 Oracle Wallet Manager Help Does Not Work

The Oracle Wallet Manager screen images and icons do not display clearly, and Oracle Wallet Manager help does not work.

Workaround:

Add owm-images.jar and owm_help.jar to the CLASSPATH variable in the owm.cl file so that the screen is displayed clearly and the Oracle Wallet Manager help works correctly.

Perform the following steps:

1. In the Start menu, right-click the OWM icon and select Properties. This gives you the location of the owm.cl file.

2. Open the owm.cl file using a text editor.

3. Add the owm-images.jar and owm_help.jar files to the CLASSPATH variable in the owm.cl file.

   Note:

   The path of the.jar should be same as of owm-3_0.jar. Typically, the owm-images.jar and owm_help.jar files are located in C:\app\user-name\product\11.2.0\ORACLE_HOME\owm\jlib\.

4. Save the changes to the owm.cl file.

This issue is tracked with Oracle bug 7506415.

4.3.2 Delete Node Deployment Procedure Not Supported

Deleting nodes using Oracle Enterprise Manager Database Control is currently not supported. However, you can manually delete the nodes. See the following for more information:

• Chapter 3, "Cloning Oracle Clusterware to Create a Cluster" and Chapter 4, "Adding and Deleting Cluster Nodes on Windows Systems" in the Oracle Clusterware Administration and Deployment Guide

• Chapter 10, "Adding and Deleting Nodes and Instances on Windows Systems" in the Oracle Real Application Clusters Administration and Deployment Guide

4.3.3 Oracle Database Extensions for .NET Support on Windows x64

If you run a .NET Stored Procedure with .NET Framework 2.0 SP2, then you get an access violation in extproc.exe. This does not occur if you run a .NET Stored Procedure with .NET Framework 2.0 SP1.

Workaround:

Install Microsoft hotfix KB974168.

4.3.4 Oracle RAC Patch Not Supported

Oracle RAC patch from Oracle Enterprise Manager Database Control is not supported.

Workaround:

Run OPatch to apply patches to Oracle RAC.

This issue is tracked with Oracle bug 9128255.

4.3.5 Help File Shortcut Not Working in Windows Server 2008

Starting Oracle Objects for OLE C++ Class Library Help from the Start menu shortcut does not open the help page in Windows Server 2008.

Workaround:

Complete the following steps to resolve this issue:

1. Right-click the shortcut.

2. Select Open With and click Internet Explorer.

3. Select the option to open the help page.

This issue is tracked with Oracle bug 8001859.

4.3.6 Incorrect Reverse Ping Command from Oracle Management Service

Oracle Management Service provides incorrect reverse ping command with Oracle Enterprise Manager Database Control. The emoms.properties file shows:

emdrep.ping.pingCommand=ping <hostname>

Workaround:

Change the property in the emoms.properties file to:

ping -n 3 -w 30000 <hostname>

This issue is tracked with Oracle bug 9211800.

4.3.7 Installing Online Patches on Windows Systems

The Opatch tool shipped with Oracle Database 11g Release 2 fails while applying online patches on Windows systems.

Workaround:

Opatch tool version 11.2.0.1.1 is available on My Oracle Support as Patch 6880880 for Oracle Database 11g Release 2. Download and install Opatch Tool version 11.2.0.1.1 before applying online patches on Windows systems.

To list the database instances patched by an online patch, run the opatch lsinventory -details command. Ensure that you run the opatch rollback command instead of opatch disableOnlinepatch command to disable and remove an online patch from the final database instance in a node. To remove the online patch that is in a disabled state, run the disableOnlinepatch command. To remove the online patch in a disabled state, use -no_sysmod flag with opatch rollback command, only when opatch lsinv -details does not list any instances for that patch. For example, use opatch rollback -id 1234 -no_sysmod, only when opatch lsinv -details show no instances patched for bug 1234.

This issue is tracked with Oracle bugs 8843060 and 9440270.

4.3.8 File Systems Mounted with Drive Letters on Oracle ADVM Volumes Are Not Visible in Windows Explorer

When the Oracle ACFS mount registry resource or the Oracle ACFS individual file system resource mounts an Oracle ADVM volume on a drive letter, the drive is not visible in Windows Explorer. This only affects the visibility of the drive letter in Windows Explorer; the volume is mounted and fully accessible using the command line, and is visible using the acfsutil.exe utility.

Workaround:

You must log out of your Windows session and log back in to view the driver letter in Windows Explorer.

This issue is tracked with Oracle bug 9343876.

4.3.9 Download Microsoft Handle Utility When Using Oracle ACFS

The Microsoft handle utility is required for proper operation of Oracle ACFS High Availability resources managed by Oracle Clusterware. Oracle Clusterware resources may not shut down properly if this command is not included.

Workaround:

Download the Microsoft Handle utility from:

http://technet.microsoft.com/en-us/sysinternals/bb896655.aspx

After downloading, place the Handle utility in the BIN directory of your Grid home.

This issue is tracked with Oracle bug 9349433.

4.3.10 Oracle Grid Infrastructure Listeners Issue

After installing Oracle Grid Infrastructure, the default listener does not listen on public addresses. As a result, even though the listener is available, the listener does not service the database or Oracle ASM instances.

Workaround 1:

If you are not using GNS with DHCP, then perform the following steps:

1. Add the local IP address and host name information to %Systemroot%\System32\Drivers\Etc\Hosts.

   For example, IP address host name such as 192.0.2.254 dbhost.

2. Restart the Oracle Grid Infrastructure listener.

This issue is tracked with Oracle bug 9382523.

Workaround 2:

If you have configured GNS and use DHCP to assign the addresses for the cluster nodes, then configure DHCP so that the IP address assigned to the host name for each cluster node (the public IP address) has a smaller numeric value than the IP addresses assigned for the Virtual IP (VIP) address, GNS VIP address, or SCAN VIP address for a node.

For example, you could use IP addresses in the range of xxx.xxx.xxx.0 – xxx.xxx.xxx.099 for only host IP addresses, and all other IP address assignments must use a value of xxx.xxx.xxx.100 or higher.

This issue is tracked with Oracle bug 9441288.

4.3.11 Shared Oracle Home Shortcuts Do Not Work on Windows Server 2008

Starting Oracle tools such as Net Configuration Assistant, Database Configuration Assistant, Database Upgrade Assistant, Net Manager, Oracle Universal Installer, and so on, from a shared Oracle home shortcut gives the following error:

Missing shortcut: windows is searching launch.exe

After displaying the error message, the shortcut exits.

Workaround:

Start the tools from a Windows command prompt with Administrator privileges.

This issue is tracked with Oracle bug 7525053.

4.3.12 addNode.bat Script Shows Success Even When One or More Node Additions Fail

When running the addNode.bat script to add one or more nodes to your cluster, if there was a failure during any installation on any of the new nodes, they are removed from the list of nodes to be adding to the cluster. As a result the updatenodelist command called after the node addition process returns a status of success even if there was a failure in adding a node.

Workaround:

None. Verify that the new nodes are correctly added to the node list after the addNode procedure. If not, check the addNode logs for any error messages. Correct the problem and repeat the addNode procedure.

This issue is tracked with Oracle bug 9481472.

4.3.13 Inventory Issue with addNode.bat Script

After running the addNode.bat script, the inventory on the new nodes does not reflect the newly added home.

Workaround:

On the node being added by the addNode.bat script, do the following:

CRS or Grid Home

cd Grid_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -cfs -local 

Database or Oracle RAC Home

cd RAC_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -cfs -local 

This issue is tracked with Oracle bugs 9544552 and 9541191.

4.3.14 Data Pump Worker Processes Do Not Run on Different Oracle RAC Instances

When running a Data Pump job in an Oracle RAC environment, if you specify a value greater than 1, then Data Pump should use the Oracle RAC resources and start additional worker processes on other nodes in the cluster up to the value of the parallel parameter. However, for Windows systems, all the Data Pump processes run on the instance where the job is created; worker processes on other nodes are not started.

Workaround:

None.

This issue is tracked with Oracle bug 9448829.

4.3.15 ORA_DBA Group Not Created By addNode.bat Script

When you use the addNode.bat script to add a node to your cluster, the operating system group ORA_DBA is not created on the nodes being added to the cluster or the user running the addNode.bat script is not added to the ORA_DBA group on the nodes being added to the cluster. This causes the root.bat script to fail.

Workarounds:

1. If you have not yet run the root.bat script as prompted by addNode.bat, then you can manually create the ORA_DBA operating system group on the nodes being added to the cluster and add the user that is running the addNode.bat script to that ORA_DBA group.

   When installing software in a cluster or adding a node to a cluster, the same operating system user performing the installation on the first node must exist on each node, and the user must be explicitly declared as a member of the local Administrators group and the ORA_DBA group on each node in the cluster.

   Alternatively, you can run the following command from Oracle home in which Oracle Clusterware is installed as the user that runs addNode.bat:

   ORACLE_BASE\ORACLE_HOME\bin\oradim.exe -ex network useradd ora_dba
   %USERDOMAIN%\%USERNAME% "Oracle DBA Group"
   

2. If you have already run the root.bat script and it failed, then perform the following steps:

   1. Deconfigure Oracle Clusterware on the nodes being added to the cluster:

      cd ORACLE_BASE\ORACLE_HOME\crs\install
      perl rootcrs.pl -deconfig -force
      

   2. Perform the workaround 1 and rerun the root.bat script.

This issue is tracked with Oracle bug 9525501.

4.3.16 ODP.NET Help Not Installed on 64-Bit Windows

If you install 32-bit Oracle software on a 64-bit Windows computer, then help for ODP.NET is not installed.

Workaround:

Use the .NET Help Setup Wizard to manually install the dynamic help. To start the wizard on an English locale, run the DataProviderHelp_en.msi file from the ORACLE_BASE\ORACLE_HOME\ODP.NET\help directory. For the Japanese locale, run the DataProviderHelp_ja.msi file from the same directory.

This issue is tracked with Oracle bug 9524039.

4.3.17 Using Oracle ACFS for the Database Home and LOB Files

If you place the Oracle Database home on Oracle ACFS, and the database uses LOB files, which are also stored on Oracle ACFS, then the Oracle ACFS home must be of the format, DRIVE_LETTER:\directory. For example:

g:\dir
g:\dir1\dir2

Otherwise, LOB related errors are shown when accessing the LOB (as in the sample schema for a basic installation). For example:

ERROR at line 1:
ORA-22288: file or LOB operation FILEOPEN failed
ORA-06512: at "SYS.DBMS_LOB", line 744
ORA-06512: at "XDB.DBMS_XDB", line 301
ORA-06512: at line 4

See "File Systems Mounted with Drive Letters on Oracle ADVM Volumes Are Not Visible in Windows Explorer" for related information. If you are not using a drive letter for the Oracle ACFS home, then several LOB related error messages appear during sample schema install due to this issue.

Workaround:

Click Ignore these errors to proceed, or click Abort to stop the database creation.

This issue is tracked with Oracle bug 9527303.

4.3.18 Oracle ACFS Recommendation for Oracle Grid Infrastructure Shutdown

Under certain conditions, errors may be issued during an Oracle Grid Infrastructure shutdown relating to activity on Oracle ACFS file systems:

acfsdismount: ACFS-04007: dismount failed due to open files on the volume

Workaround:

To avoid these errors, unmount all Oracle ACFS file systems manually before Oracle Grid Infrastructure shutdown. You can use the Windows handle and wmic commands to identify processes which are active on the file systems. Ensure that these processes are no longer active, unmount all ACFS file systems, then reissue Oracle Grid Infrastructure shutdown.

This issue is tracked with Oracle bugs 9539710 and 9445489.

4.3.19 Oracle Notification Service Fails to Stop During Oracle Grid Infrastructure Shutdown

You might encounter the following error during shutdown of Oracle Grid Infrastructure on a node:

CRS-5016 (or CRS-2675): stop action failed for ora.eons resource

Workaround:

Perform the following steps to manually stop the ora.eons resource and shut down Oracle Grid Infrastructure:

1. Disable the ora.eons resource on the node by running the following command:

   crsctl modify resource ora.eons -attr "ENABLED@SERVERNAME(nodename)=0"
   

2. Find and stop the java process for the ora.eons resource, oracle.supercluster.impl.cluster.EONSServerImpl.

   1. To determine the process ID (PID) of the process, you can use a utility such as the Microsoft Windows Management Utility. For example:

      wmic RETURN process
      

      While logged in as an Administrator user, start the Microsoft Windows Management Utility using the following command:

      C:\> wmic
      

      When the utility prompt appears, run the process command, as shown here:

      wmic:root\cli> process
      

      A long list of processes and the command used to start those processes is displayed. For the processes named java.exe, find the one associated with the eons process. It will appear similar to the following, where Grid_home is the directory in which Oracle Grid Infrastructure is installed on the local node:

      java.exe   ”Grid_home\jdk\jre\bin\java\” –Doracle.supercluster.cluster.server=eonsd –classpath …
      

   2. To get the PID for the process, use the following command:

      wmic:root\cli>  process where (commandline like ”%eons%”) list brief
      

      This command returns output similar to the following:

      HandleCount  Name      Priority  ProcessId  ThreadCount  WorkingSetSize
      39           cmd.exe   8         1292       1            2654208
      705          java.exe  8         5540       38           54411264
      

      The PID is listed in the ProcessID column for the java.exe process.

   3. Exit the Microsoft Windows Management Utility using the following command:

      wmic:root\cli> exit
      

   4. Stop the java process, using a command similar to the following, where PID is the process ID obtained in the previous step:

      C:\> kill –f PID
      

3. Stop Oracle Grid Infrastructure on the node by running the following command, where nodename is the name of the node you are attempting to shut down:

   C:\> crsctl stop cluster -n nodename
   

   Ignore any error messages displayed on the console while Oracle Grid Infrastructure is being shut down.

4. Restart Oracle Grid Infrastructure by running the following command, where nodename is the name of the node:

   C:\> crsctl start cluster -n nodename
   

5. Enable the ora.eons resource by running the following command:

   crsctl modify resource ora.eons -attr "ENABLED@SERVERNAME(nodename)=1"
   

6. Start the ora.eons resource by running the command:

   crsctl start res ora.eons -n nodename
   

This issue is tracked with Oracle bug 9524640.

4.3.20 Global Services Daemon (GSD) Resource Shows ONLINE After Oracle Grid Infrastructure Restart

The GSD resource should stay OFFLINE in a cluster environment when there are no Oracle 9i databases configured. After you restart Oracle Clusterware, it is possible the GSD resource status could be changed to ONLINE. There is no issue with keeping the GSD resource ONLINE, but if this resource is not needed (there are no Oracle 9i databases to be managed) and you want to keep this resource OFFLINE, then use the suggested workaround.

Workaround:

Use the following command to stop GSD on all the nodes:

srvctl stop nodeapps -g

Also, ensure that the file, Grid_home\admin\gsd.pid, does not exist on any of the cluster nodes.

This issue is tracked with Oracle bug 9489026.

4.3.21 VIP Does Not Fail Back When Public Network Interface Card (NIC) Status Changes Rapidly

If the Public NIC is disabled and enabled within a short span of time, then the VIP resources do not fail back to the node where the NIC was enabled.

Workaround:

Use CRSCTL to relocate the node VIP:

crsctl relocate res ora.nodename.vip -n nodename
 

This issue is tracked with Oracle bug 9369822.

4.3.22 cluvfy.bat Fails to Check Shared Resource Status for Node Addition

When you use cluvfy.bat to validate the new node to be added to a cluster with OCFS-based OCR and voting disks, cluvfy.bat displays an error message stating that the resources are not shared.

Workaround:

You can ignore this error message.

This issue is tracked with Oracle bug 9546850.

5.2.1 Oracle ACFS Drivers Installation Error

Oracle ASM Configuration Assistant (ASMCA) may report that the Oracle ASM Cluster File System (ACFS) drivers are not installed when in reality the drivers are installed but not loaded. This is seen in the Oracle ASM Instances tab of ASMCA user interface after Oracle ASM is configured. There is no functionality impact and you can check if the drivers are installed by running grid_home\bin\acfsdriverstate installed. This utility prints the correct information.

Workaround:

None.

This issue is tracked with Oracle bug 9948257.

5.2.2 Oracle Configuration Manager Service Not Removed on Remote Node When Performing Cluster Deinstallation

When removing Oracle Configuration Manager from an Oracle home, in some cases, the service is stopped on remote nodes but not removed.

Workaround:

Remove the service using the sc command:

sc remove OracleORACLE_HOME_IDENTIFIERConfigurationManager

This issue is tracked with Oracle bug 10074102.

5.2.3 Service Not Removed When Deleting a Node on Shared Oracle RAC

When deleting a node on shared Oracle RAC home, OraClrAgnt service is stopped but not removed.

Workaround:

Manually remove the service using the sc command:

sc remove OracleORACLE_HOME_NAMEClrAgent

This issue is tracked with Oracle bug 10074070.

5.2.4 Issues Associated with Oracle SID Containing Underscore Character

Oracle Database 11.2 supports Oracle SID having an underscore but it is not compatible with the earlier client releases with a bequeath connection. The issue affects any Oracle Database Client 11.1.0.7 or earlier client, when connecting to the database over a bequeath connection. This includes all OCI clients and Oracle Enterprise Manager agents that monitor the database.

Workaround:

Use only alphanumeric characters including hyphens for Oracle SID or connect to the database using TCP connection.

This issue is tracked with Oracle bug 7611024.

5.2.5 Oracle Clusterware Upgrade Error

When you use Oracle Universal Installer to upgrade Oracle Clusterware 11g Release 1 (11.1.0.6), or Oracle Clusterware 11g Release 1 (11.1.0.7) to Oracle Clusterware 11g Release 2 (11.2), you may receive the following error:

[INS-40406] There is no previous version of grid software detected on the system.

Workaround:

If you are performing the upgrade on a remote node, a node other than the one on which Oracle Clusterware 11g Release 1 (11.1.0.6) was installed, then the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc may not have been set on that node.

Ensure that you manually set the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc to the location of the Oracle Inventory directory of the node on which you are performing the upgrade installation.

This issue is tracked with Oracle bug 9210861.

5.2.6 Remote Agents Must Be Manually Reloaded After Cluster Ready Services Upgrade and Oracle ASM Upgrade

When attempting to upgrade Cluster Ready Services or Oracle ASM to 11.2 release using the Automatic Storage Management Configuration Assistant, the upgrade succeeds, but the configuration changes made to the Agents on a remote host fail to reload. As a result, Oracle Enterprise Manager Grid Control and Database Control cannot monitor the Oracle ASM and Cluster Ready Services targets.

Workaround:

Manually reload the remote agents after the upgrade using the following command:

emctl reload agent

This issue is tracked with Oracle bug 9398380.

5.2.7 Deinstallation Tool Does Not Remove Oracle Home

When using the Deinstallation Tool to remove an Oracle home that has Oracle Enterprise Manager Database Control configured and that was installed on Oracle ACFS, the ORACLE_BASE\ORACLE_HOME\host_sid\sysman directory and its sub directories cannot be deleted, preventing the Oracle home from being removed.

Workaround:

Before performing these operations, run the following command as the Administrator user:

cacls ORACLE_BASE\ORACLE_HOME\HOSTNAME_DBNAME\sysman\emd\upload\hsperfdata_SYSTEM /E /P %USERDOMAIN%\%USERNAME%:F 

This issue is tracked with Oracle bug 9444816.

5.2.8 Deinstallation Tool Fails to Remove Registry Entries

When deinstalling a single instance Oracle Database, Oracle Database Client, or Oracle RAC (only 64-bit Windows servers), some registry entries are not removed.

Workaround:

• Manually remove the registry entry, OracleDBConsoleSID:

  HKLM\System\CurrentControlSet\Services\EventLog\Application
   
  

  This issue is tracked with Oracle bug 9523875.

• Manually remove the registry entries, Oracle.SID and Oracle.VSSWriter.SID:

  HKLM\System\CurrentControlSet\Services\EventLog\Application
   
  

  This issue is tracked with Oracle bug 9523866.

• Manually remove the following registry entries:

  HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\ODP.NET
  HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\_DosLinks
  HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE
  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Oracle11
  

  This issue is tracked with Oracle bug 10043825.

5.2.9 Root Script Status Not Displayed with Oracle Grid Infrastructure

When running rootcrs.pl or roothas.pl to configure, upgrade, or deconfigure Oracle Grid Infrastructure, the script does not display the success or failure status in the command window.

Workaround:

None. Check the log file generated at the following location to know the status of the script:

Grid_home/cfgtoollogs/rootcrs_host_name.log 

This issue is tracked with Oracle bug 9473262.

5.2.10 Time Issue with Deinstallation Tool to Remove Oracle RAC Home

When using the Deinstallation Tool to remove an Oracle RAC home that was installed on Oracle ACFS, the tool takes a longer amount of time than expected to finish.

Workaround:

None. Wait an additional 20-30 minutes for the deinstallation tool to finish.

This issue is tracked with Oracle bug 9478372.

5.2.11 Oracle MTS Recovery Service Issue When Installing Oracle RAC

In an Oracle RAC installation, Oracle MTS Recovery Service on the remote nodes does not have the correct host name registered for string variable HOST under registry key HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0. This variable contains the name of the host from where the installation was performed instead of the name of the local host. This registry variable is used by COM+ applications to pass information to Oracle databases so that they can connect back to the Windows computer if the database must resolve in-doubt COM+ transactions. The incorrect host name that is registered can cause transaction recovery issues.

Workaround:

To fix this, use Registry Editor (regedit) on the remote Windows system and look under the registry key:

HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0

The string-type values for the following key must be set correctly:

Name:            *Host*
Type:            REG_SZ
Data: << the name of the host-machine e.g. myhost@example.com <mailto:myhost@example.com>>> 

This key must contain the host name of the system on which the registry key exists.

This issue is tracked with Oracle bug 9483665.

5.2.12 Deinstallation Tool Fails to Remove Oracle Grid Infrastructure for a Cluster in Silent Mode

When using the Deinstallation Tool to remove Oracle Grid Infrastructure for a cluster in silent mode, the tool fails with the following error message:

ERROR:  As the privileged user, Execute \deinstall\crsdc.bat on every
node  in this cluster except for this node, and then execute
\deinstall\rootdeinstall.bat on this node to remove the configuration of the
Oracle Clusterware and to completely remove the Oracle software 

Workaround:

Follow these steps when the deinstallation fails:

1. Run crsdc.bat on all the nodes as instructed by the Deinstallation Tool.

2. Run rootdeinstall.bat from the deinstall directory on node 1.

This issue is tracked with Oracle bug 9540713.

5.2.13 Using Deinstallation Tool for a Single-Instance Oracle Database Returns Errors

Removing Oracle home fails with file in use errors as the services and processes running from the Oracle home are not removed completely by Microsoft's Service Control Manager before the Oracle home removal.

Workaround:

Remove the Oracle home manually after the deinstall tool exits.

This issue is tracked with Oracle bug 10035858.

5.2.14 Deinstallation Tool Fails to Remove Shared Oracle RAC Home

When using the Deinstallation tool, deinstall.bat, to remove a shared Oracle RAC home, you get the error message, Failed to delete directory.., for the Oracle home being removed.

Workaround:

Use Process Monitor, PMON, to locate and stop the processes running out of the Oracle home being removed. Then remove the directory.

This issue is tracked with Oracle bug 10145215.

5.2.15 Deinstallation Tool Does Not Remove invDetails.properties File

The Deinstallation Tool fails to remove the invDetails.properties file when deinstalling Oracle RAC home.

Workaround:

Manually remove the dbhome\inventory directory.

This issue is tracked with Oracle bug 9543109.

5.2.16 Issues While Installing Oracle Database with Oracle ASM

If you install Oracle Database release 11.1 or earlier and during installation choose to create a database with storage configured on Oracle ASM release 11.2.0.2, then the installation fails while the configuration tools are running, specifically Database Configuration Assistant.

Workaround:

Install the Oracle Database release 11.1 or earlier software and choose the software only installation option. After the installation completes, create the single-instance database by launching Database Configuration Assistant from the newly installed Oracle home. For example, run the following command from a command prompt:

ORACLE_BASE\ORACLE_HOME\bin\dbca.bat

This issue is tracked with Oracle bug 10149589.

5.2.17 Services for an Oracle RAC 10.2 Database Instance Fail to Start

If you run an Oracle Database release 10.2 or 11.1 on Oracle Grid Infrastructure release 11.2, after a node is restarted, the database instance resource running on that node may erroneously report an online state when it is actually offline. Any service associated with the instance fails to start.

Workaround:

Manually issue a start against the instance by running the following commands on the node that was restarted:

srvctl stop instance -d db_unique_name -i instance_name
srvctl start instance -d db_unique_name -i instance_name

This issue is tracked with Oracle bug 10224187.

5.2.18 rootcrs_node.log File Fails When Installing Oracle Clusterware on German Windows Server x64

When installing Oracle Clusterware on a German operating system, the rootcrs_node.log file fails with the following error message:

Stop of resource "ora.asm -init" failed

Workaround:

When Oracle Clusterware installation fails with this error, perform the following steps:

1. Check the log file for the following error message:

   Stop of resource "ora.asm -init" failed
   

2. Apply the following command on each node:

   "crsctl modify resource ora.asm -attr CHECK_TIMEOUT=60 -init"
   

3. Click retry in Oracle Universal Installer (OUI) window to continue with the Oracle Clusterware installation.

4. If the same error occurs, repeat step 2.

5. After a successful Oracle Clusterware installation, check the CRS status:

   "crsctl stat res -t" 
   

6. If any service is offline, start the service.

   For example, if ora.DATA.dg service on node2 is offline, run the following command to start the service:

   "srvctl start diskgroup -g data -n node2"
   

This issue is tracked with Oracle bug 10250914.

5.2.19 Database Configuration Assistant Does Not Run When Installing Oracle ACFS

The GetVolumePathname Microsoft API on Windows Server 2008 R2 shows incorrect results when installing an Oracle ACFS file system hosted on an Oracle ADVM volume. As a result, Database Configuration Assistant and other configuration tools cannot run during the Oracle ACFS file system install.

Workaround:

Run Database Configuration Assistant manually after installing Oracle ACFS on Windows Server 2008 R2.

This issue is tracked with Oracle bug 10242246.

5.2.20 Issues When Upgrading Oracle Grid Infrastructure

When upgrading Oracle Grid Infrastructure from release 11.2.0.1 to release 11.2.0.2, you might encounter an issue when Oracle Grid Infrastructure Home column on the Oracle Grid Infrastructure Node Selection screen reflects an incorrect empty value.

Workaround:

Follow these steps to resume Oracle Grid Infrastructure upgrade:

1. Stop OracleRemExecService by running the following command:

   net stop OracleRemExecService
   

2. To check if OracleRemExecService has stopped, run the following command:

   sc query OracleRemExecService
   

   The following error messages appear after running the command:

   sc EnumQueryServicesStatus:OpenService FAILED 1060:
   The specified service does not exist as an installed service. 
   

3. Run the following command to re-create the service:

   11.2.0.2 Grid shiphome\runcluvfy stage -pre crsinst -n node-list
   

   where node-list is the list of nodes for Grid infrastructure. Ignore the errors.

4. Rerun the asmca utility to upgrade the Oracle Grid Infrastructure.

This issue is tracked with Oracle bug 10318692.

5.2.21 Deinstallation Tool Might Not Remove ORACLECLUSTERVOLUMESERVICE

When using the Deinstallation Tool to remove a Clusterware home, you might encounter the issue that removing service, OracleClusterVolumeService, failed.

Workaround:

Manually remove the service, OracleClusterVolumeService.

This issue is tracked with Oracle bug 9405729.

5.2.22 Deinstallation Issues When Removing Oracle Home in Sequence

After removing a database Oracle home or a database client Oracle home, the deinstallation might hang if you try to remove another Oracle home, without any new Oracle home installed in between the two deinstallations.

Workaround:

Stop the reg.exe process from the Task Manager application to continue the deinstallation of the Oracle home.

This issue is tracked with Oracle bug 10354341.

5.3.1 Help File Shortcut Not Working in Windows Server 2008

Starting Oracle Objects for OLE C++ Class Library Help from the Start menu shortcut does not open the help page in Windows Server 2008.

Workaround:

Complete the following steps to resolve this issue:

1. Right-click the shortcut.

2. Select Open With and click Internet Explorer.

3. Select the option to open the help page.

This issue is tracked with Oracle bug 8001859.

5.3.2 Memory Mapped File Limitations with Oracle ACFS on Windows Server 2008

Due to limitations in the Windows Server 2008 operating system, Oracle ACFS on Windows Server 2008 has certain limitations with file changes made through memory mapped access. If file changes are made through memory mapped access on one node, then those changes may not be seen on other nodes until the application on the first node either explicitly flushes the view of the file, or unmaps the view, and closes the file. In addition, if an Oracle ACFS snapshot is created while modifications to files made through memory mapping are in the page cache and have not been flushed to disk, these changes are not reflected in the snapshot.

These limitations do not exist on Windows Server 2008 R2.

5.3.3 Using Oracle ACFS for the Database Home and LOB Files

If you place the Oracle Database home on Oracle ACFS, and the database uses LOB files, which are also stored on Oracle ACFS, then the Oracle ACFS home must be of the format, DRIVE_LETTER:\directory. For example:

g:\dir
g:\dir1\dir2

Otherwise, LOB related errors are shown when accessing the LOB (as in the sample schema for a basic installation). For example:

ERROR at line 1:
ORA-22288: file or LOB operation FILEOPEN failed
ORA-06512: at "SYS.DBMS_LOB", line 744
ORA-06512: at "XDB.DBMS_XDB", line 301
ORA-06512: at line 4

See "File Systems Mounted with Drive Letters on Oracle ADVM Volumes Are Not Visible in Windows Explorer" for related information. If you are not using a drive letter for the Oracle ACFS home, then several LOB related error messages appear during sample schema install due to this issue.

Workaround:

Click Ignore these errors to proceed, or click Abort to stop the database creation.

This issue is tracked with Oracle bug 9527303.

5.3.4 Oracle Database Extensions for .NET Support on Windows x64

If you run a .NET Stored Procedure with .NET Framework 2.0 SP2, then you get an access violation in extproc.exe. This does not occur if you run a .NET Stored Procedure with .NET Framework 2.0 SP1.

Workaround:

Install Microsoft hotfix KB974168.

5.3.5 Issues While Installing .NET Related Products through Mapped Drives

When installing .NET related products such as Oracle Data Provider for .NET, Oracle Providers for ASP.NET, and so on, to an ORACLE_HOME and ORACLE_BASE that are located on a mapped drive on either 32-bit or 64-bit operating system, you may encounter a System.IO.FileLoadException exception indicating that .NET Framework failed to load OraProvCfg.exe due to a failure in granting minimum permission requests.

Workaround:

Install .NET Framework 2.0 SP2 or later for the appropriate operating system.

This issue is tracked with Oracle bug 10225338.

5.3.6 Issues While Specifying the Software Updates Download Location

Oracle Universal Installer does not support spaces and double quotation marks in the path of the download location when the Software Updates option is chosen.

This issue is tracked with Oracle bug 10194635.

5.3.7 Shared Oracle Home Shortcuts Do Not Work on Windows Server 2008

Starting Oracle tools such as Net Configuration Assistant, Database Configuration Assistant, Database Upgrade Assistant, Net Manager, Oracle Universal Installer, and so on, from a shared Oracle home shortcut gives the following error:

Missing shortcut: windows is searching launch.exe

After displaying the error message, the shortcut exits.

Workaround:

Start the tools from a Windows command prompt with Administrator privileges.

This issue is tracked with Oracle bug 7525053.

5.3.8 Download Microsoft Handle Utility When Using Oracle ACFS

The Microsoft handle utility is required for proper operation of Oracle ACFS High Availability resources managed by Oracle Clusterware. Oracle Clusterware resources may not shut down properly if this command is not included.

Workaround:

Download the Microsoft Handle utility from:

http://technet.microsoft.com/en-us/sysinternals/bb896655.aspx

After downloading, place the Handle utility in the BIN directory of your Grid home.

This issue is tracked with Oracle bug 9349433.

5.3.9 Oracle Grid Infrastructure Listeners Issue

After installing Oracle Grid Infrastructure, the default listener does not listen on public addresses. As a result, even though the listener is available, the listener does not service the database or Oracle ASM instances.

Workaround 1:

If you are not using GNS with DHCP, then perform the following steps:

1. Add the local IP address and host name information to %Systemroot%\System32\Drivers\Etc\Hosts.

   For example, IP address host name such as 192.0.2.254 dbhost.

2. Restart the Oracle Grid Infrastructure listener.

This issue is tracked with Oracle bug 9382523.

Workaround 2:

If you have configured GNS and use DHCP to assign the addresses for the cluster nodes, then configure DHCP so that the IP address assigned to the host name for each cluster node (the public IP address) has a smaller numeric value than the IP addresses assigned for the Virtual IP (VIP) address, GNS VIP address, or SCAN VIP address for a node.

For example, you could use IP addresses in the range of xxx.xxx.xxx.0 – xxx.xxx.xxx.099 for only host IP addresses, and all other IP address assignments must use a value of xxx.xxx.xxx.100 or higher.

This issue is tracked with Oracle bug 9441288.

5.3.10 Installing Online Patches on Windows Systems

The Opatch tool shipped with Oracle Database 11g Release 2 fails while applying online patches on Windows systems.

Workaround:

Opatch tool version 11.2.0.1.1 is available on My Oracle Support as Patch 6880880 for Oracle Database 11g Release 2. Download and install Opatch Tool version 11.2.0.1.1 before applying online patches on Windows systems.

To list the database instances patched by an online patch, run the opatch lsinventory -details command. Ensure that you run the opatch rollback command instead of opatch disableOnlinepatch command to disable and remove an online patch from the final database instance in a node. To remove the online patch that is in a disabled state, run the disableOnlinepatch command. To remove the online patch in a disabled state, use -no_sysmod flag with opatch rollback command, only when opatch lsinv -details does not list any instances for that patch. For example, use opatch rollback -id 1234 -no_sysmod, only when opatch lsinv -details show no instances patched for bug 1234.

This issue is tracked with Oracle bugs 8843060 and 9440270.

5.3.11 Inventory Issue with addNode.bat Script

After running the addNode.bat script, the inventory on the new nodes does not reflect the newly added home.

Workaround:

On the node being added by the addNode.bat script, do the following:

CRS or Grid Home

cd Grid_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -cfs -local 

Database or Oracle RAC Home

cd RAC_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -cfs -local 

This issue is tracked with Oracle bugs 9544552 and 9541191.

5.3.12 Issues While Deleting Oracle RAC on a Remote Node

When using Oracle Database Configuration Assistant to delete an Oracle RAC instance on a remote node, you may encounter the following error:

EMCA failed due to the following error - Error updating EM configuration for node node_name

Workaround:

Use Process Monitor, PMON, to identify and stop ohFilesPerm.pl perl process. Then remove the Oracle Enterprise Manager Database Control state directory:

ORACLE_HOME/nodename_db_unique_name

This issue is tracked with Oracle bug 10073887.

5.3.13 SQL Developer Shortcut Appears After Oracle Database Client Upgrade

After upgrading the Oracle Database Client release 11.2.0.1 to release 11.2.0.2, the SQL Developer shortcut appears.

Workaround:

Do not use this SQL Developer shortcut as it is not supported.

This issue is tracked with Oracle bug 10078387.

5.3.14 Issues While Relocating Grid Naming Service on Windows Server 2008

After installing Oracle Grid Infrastructure with Grid Naming Service enabled on Windows Server 2008, when you relocate the Grid Naming Service, the name resolution fails.

Workaround:

If the Grid Naming Service is relocated, then run the following command from the previous Grid Naming Service node:

ping gns-vip

This issue is tracked with Oracle bug 10124639.

5.3.15 Shortcuts for Oracle Software Not Removed from Start Menu When Removing a Node

Start menu shortcut entries are not removed when a node is deleted from a shared Oracle RAC home.

Workaround:

Manually remove the Start menu shortcut entries:

1. From the Start menu, select All Programs, then right-click Oracle-HOME_NAME.

2. Select the delete option and confirm the deletion.

This issue is tracked with Oracle bug 10074059.

5.3.16 PATH Variable Not Updated on Recently Deleted Node that Uses a Shared Oracle Home

If your Oracle RAC uses a shared Oracle home and you delete a node, then the PATH environment variable on the node being deleted is not updated to remove the location of the Oracle home.

Workaround:

Update the PATH variable manually on deleted node using the following steps:

1. From the Start menu, select Control Panel, then open System.

2. Select Advanced and click Environment Variables.

3. Select PATH variable from System Variables and double-click to get variable value.

4. Remove Oracle RAC home reference from variable value and click OK.

5. Click OK on Environment Variables dialog and then click Apply on System Properties dialog.

This issue is tracked with Oracle bug 10074038.

5.3.17 Access Violation Issues with ODP.NET Using Oracle Streams Advanced Queuing

ODP.NET applications using Oracle Streams Advanced Queuing with XML as the message payload may encounter access violations and possible hangs during .NET garbage collection.

Workaround:

Call the Dispose() method on the OracleAQQueue object when the application is done using it.

This issue is tracked with Oracle bug 10389639.

6.2.1 Issues While Installing Oracle Database with Oracle ASM on Windows x64

If you install Oracle Database release 11.1 or earlier and during installation choose to create a database with storage configured on Oracle ASM release 11.2.0.3, then the installation fails while the configuration tools are running, specifically Database Configuration Assistant.

Workaround:

Install the Oracle Database release 11.1 or earlier software and choose the software only installation option. After the installation completes, create the single-instance database by launching Database Configuration Assistant from the newly installed Oracle home. For example, run the following command from a command prompt:

ORACLE_BASE\ORACLE_HOME\bin\dbca.bat

This issue is tracked with Oracle bug 10149589.

6.2.2 Deinstallation Tool Fails to Remove Shared Oracle RAC Home on Windows x64

When using the Deinstallation tool, deinstall.bat, to remove a shared Oracle RAC home, you get the error message, Failed to delete directory.., for the Oracle home being removed.

Workaround:

Use Process Monitor, PMON, to locate and stop the processes running out of the Oracle home being removed. Then remove the directory.

This issue is tracked with Oracle bug 10145215.

6.2.3 Oracle Configuration Manager Service Not Removed on Remote Node When Performing Cluster Deinstallation on Windows x64

When removing Oracle Configuration Manager from an Oracle home, in some cases, the service is stopped on remote nodes but not removed.

Workaround:

Remove the service using the sc command:

sc remove OracleORACLE_HOME_IDENTIFIERConfigurationManager

This issue is tracked with Oracle bug 10074102.

6.2.4 Issues While Installing Oracle RAC with Oracle ASM on Windows x64

If you install Oracle RAC release 11.2.0.1 with Oracle ASM release 11.2.0.3, then the following error may occur:

OUI-35078: Error in updating env on nodes 'node_name'.

Workaround:

Perform the following steps:

1. Run the following command from the 11.2.0.3 Grid home:

   srvctl stop cvu
   

2. Stop the Oracle Enterprise Manager Database Control using the following command:

   emctl stop agent
   

3. Stop the OracleRemExecService from all the nodes.

4. Then, install Oracle RAC 11.2.0.1 with Oracle ASM release 11.2.0.3.

This issue is tracked with Oracle bug 12720534.

6.2.5 Microsoft Visual C++ 2005 Redistributable Package Required for the Deinstallation Tool

If the system administrator removes the Microsoft Visual C++ 2005 Package after installing Oracle Database, then running the deinstallation tool shows the following error message:

OraOLEDBPus11.dll: The specified module could not be found

Workaround:

Reinstall the Microsoft Visual C++ 2005 Redistributable Package.

This issue is tracked with 12797310.

6.2.6 Deinstallation Tool Issues with Common Oracle Base Location on Windows x64

If you install a 32-bit Oracle Database and a 64-bit Oracle Database in the same ORACLE_BASE, then it may lead to unexpected results when using the deinstallation tool to remove one of the databases. The deinstallation tool removes all the Oracle home under Oracle base if these Oracle homes do not use the same central inventory.

Workaround:

Avoid using multiple central inventories. Do not use the same ORACLE_BASE for 32-bit and 64-bit database installations.

This issue is tracked with Oracle bugs 12808859 and 12864734.

6.2.7 Deinstallation Tool Fails to Remove Registry Entries

When deinstalling a single instance Oracle Database, Oracle Database Client, or Oracle RAC (only 64-bit Windows servers), some registry entries are not removed.

Workaround:

• Manually remove the registry entry, Oracle Services for MTS:

  HKLM\System\CurrentControlSet\Services\EventLog\Application\Oracle Services for MTS
   
  

  This issue is tracked with Oracle bug 9524210.

• Manually remove the following registry entries:

  HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\ODP.NET
  HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\_DosLinks
  HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE
  HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Oracle11
  

  This issue is tracked with Oracle bug 10043825.

• Manually remove the registry entry, OracleDBConsoleSID:

  HKLM\System\CurrentControlSet\Services\EventLog\Application
   
  

  This issue is tracked with Oracle bug 9523875.

6.2.8 Oracle Database Downgrade Issues on Windows x64

If you downgrade a database to Oracle Database release 11.2.0.2, then you might see the following invalid dictionary objects in the CTXSYS schema:

SQL>SELECT COMP_NAME, VERSION, STATUS FROM SYS.DBA_REGISTRY;
 
Oracle Text                         11.2.0.2.0      INVALID
 
SQL> select object_name, object_id, owner from all_objects where
 
status='INVALID';
 
OBJECT_NAME                     OBJECT_ID OWNER
 
------------------------------ ---------- ------------------------------
 
DRIMLX                              55736 CTXSYS
 
DRISPL                              55746 CTXSYS
 
DRIXMD                              55754 CTXSYS

You may also see ORA-00904 and ORA-00942 errors in the relod.log file, created by running catrelod.sql.

Workaround:

None

This issue is tracked with Oracle bug 12875131.

6.2.9 Prerequisite Check Issue with Firewall on Windows x64

While performing prerequisite checks during Oracle Grid Infrastructure installation on Windows x64, if the Windows Firewall is not disabled, then the Cluster Verification Utility displays the following warning:

PRVG-5906: Windows firewall is enabled on the node "node_list"
Cause: Windows firewall status was found to be enabled.
Action: To disable Windows firewall, add the REG_DWORD registry entry by name
'EnableFirewall' and value 0 to the
'HKEY_LOCAL_MACHINE/System/CurrentControlSet/Services/Tcpip/SharedAccess/Param
eters/FirewallPolicy/StandardProfile' and
'HKEY_LOCAL_MACHINE/System/CurrentControlSet/Services/Tcpip/SharedAccess/Param
eters/FirewallPolicy/DomainProfile' sub key. It is recommended to back up the
Windows Registry before proceeding with any changes, Restart your system to
make your changes effective after changing the registry.

Incorrect registry key to disable the firewall is specified in the message.

Workaround:

The correct registry key is HKEY_LOCAL_MACHINE/System/CurrentControlSet/Services/SharedAccess/Parameters/FirewallPolicy/StandardProfile and HKEY_LOCAL_MACHINE/System/CurrentControlSet/Services/SharedAccess/Parameters/FirewallPolicy/DomainProfile.

This issue is tracked with Oracle bug 12926023.

6.2.10 Deinstallation Tool Errors When Trying to Uninstall Oracle ACFS or Oracle ADVM on Windows x64

When you run the deinstallation tool from the ORACLE_BASE\ORACLE_HOME\deinstall directory, it returns the following errors:

The system cannot find the path specified.
Unable to uninstall ADVM/ACFS 

Oracle ACFS is not uninstalled.

Workaround:

Set the ORACLE_HOME environment variable to the location of the ORACLE_HOME you are removing before you start the deinstallation tool.

This issue is tracked with Oracle bug 13001410.

6.2.11 PRKN-1019 Error During Oracle Grid Infrastructure Upgrade to Release 11.2.0.3 on Windows x64

While upgrading Oracle Grid Infrastructure from release 11.2.0.1 to release 11.2.0.3, the following errors are reported:

OUI-35078: Error in updating env on nodes 'node_name'.
 
[PRKN-1019 : Failed to create directory
""%TMPDIR%\oraremservice\" on node
"node_name", Error: "0|The network path was not found. " .]

However, the specified directory exists on the node being upgraded.

Workaround:

Perform the following steps before restarting the Oracle Grid Infrastructure upgrade:

1. Stop the OracleRemExecService on all the nodes in a cluster.

2. Stop the dbconsole process and EMagent on all the nodes in a cluster by running the following command on each node:

   emctl stop dbconsole
   

This issue is tracked with Oracle bug 13052343.

6.2.12 Issues When Deinstallation Tool Runs in Silent Mode on Windows x64

When you run the deinstallation tool in silent mode using ORACLE_HOME\deinstall\deinstall.bat, the temporary files and folders are not cleaned up after exiting the deinstallation tool and running the rootdeinstall.bat file.

Workaround:

Manually delete the files or folders that are not required under the system level %Temp%\deinstall_timestamp folder.

This issue is tracked with Oracle bug 13025755.

6.2.13 PRKU-1058 Error When Using Deinstallation Tool to Remove Oracle Grid Infrastructure on Windows x64

When running the deinstallation tool with the -silent and -paramfile options, and the parameter file specifies that the Oracle ASM disk groups should not be dropped, the command completes but the following services and entries are not removed from the registry:

Services

• OracleOraCrs11g_home1TNSListenerLISTENER_SCAN1

• OracleOraCrs11g_home1TNSListenerLISTENER_SCAN2

• OracleOraCrs11g_home1TNSListenerLISTENER_SCAN3

• OracleRemExecService

Registry Keys

• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\OracleOraCrs11g_home1TNSListenerLISTENER_SCAN1

• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\OracleOraCrs11g_home1TNSListenerLISTENER_SCAN2

• HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\OracleOraCrs11g_home1TNSListenerLISTENER_SCAN3

Workaround:

None

This issue is tracked with Oracle bug 13074261.

6.2.14 Oracle Grid Infrastructure Release 11.2.0.3 Upgrade Issues with Oracle RAC Release 11.2.0.1

When you upgrade Oracle Grid Infrastructure to release 11.2.0.3 and start installing Oracle RAC release 11.2.0.1, then the Oracle Database Configuration Assistant shows the following error message towards the end of the database creation:

java.exe - Entry Point Not Found
 
The procedure entry point clsunsinit could not be located in the dynamic link library orahasgen11.dll.
---------------------------
OK  
---------------------------

Workaround:

Click OK and proceed with the database creation process. Log into the Oracle Database Enterprise Manager Control and navigate to the Cluster home page. Click Monitoring Configuration and update the eONS port for cluster target property.

This issue is tracked with Oracle bug 13074121.

6.2.15 Service Not Removed When Deleting a Node on Shared Oracle RAC

When deleting a node on shared Oracle RAC home, OraClrAgnt service is stopped but not removed.

Workaround:

Manually remove the service using the sc command:

sc remove OracleORACLE_HOME_NAMEClrAgent

This issue is tracked with Oracle bug 10074070.

6.2.16 Deinstallation Tool Does Not Remove invDetails.properties File

The Deinstallation Tool fails to remove the invDetails.properties file when deinstalling Oracle RAC home.

Workaround:

Manually remove the dbhome\inventory directory.

This issue is tracked with Oracle bug 9543109.

6.2.17 Using Deinstallation Tool for a Single-Instance Oracle Database Returns Errors

Removing Oracle home fails with file in use errors as the services and processes running from the Oracle home are not removed completely by Microsoft's Service Control Manager before the Oracle home removal.

Workaround:

Remove the Oracle home manually after the deinstall tool exits.

This issue is tracked with Oracle bug 10035858.

6.2.18 Remote Agents Must Be Manually Reloaded After Cluster Ready Services Upgrade and Oracle ASM Upgrade

When attempting to upgrade Cluster Ready Services or Oracle ASM to 11.2 release using the Automatic Storage Management Configuration Assistant, the upgrade succeeds, but the configuration changes made to the Agents on a remote host fail to reload. As a result, Oracle Enterprise Manager Grid Control and Database Control cannot monitor the Oracle ASM and Cluster Ready Services targets.

Workaround:

Manually reload the remote agents after the upgrade using the following command:

emctl reload agent

This issue is tracked with Oracle bug 9398380.

6.2.19 Deinstallation Tool Does Not Remove Oracle Home

When using the Deinstallation Tool to remove an Oracle home that has Oracle Enterprise Manager Database Control configured and that was installed on Oracle ACFS, the ORACLE_BASE\ORACLE_HOME\host_sid\sysman directory and its sub directories cannot be deleted, preventing the Oracle home from being removed.

Workaround:

Before performing these operations, run the following command as the Administrator user:

cacls ORACLE_BASE\ORACLE_HOME\HOSTNAME_DBNAME\sysman\emd\upload\hsperfdata_SYSTEM /E /P %USERDOMAIN%\%USERNAME%:F 

This issue is tracked with Oracle bug 9444816.

6.2.20 Oracle MTS Recovery Service Issue When Installing Oracle RAC

In an Oracle RAC installation, Oracle MTS Recovery Service on the remote nodes does not have the correct host name registered for string variable HOST under registry key HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0. This variable contains the name of the host from where the installation was performed instead of the name of the local host. This registry variable is used by COM+ applications to pass information to Oracle databases so that they can connect back to the Windows computer if the database must resolve in-doubt COM+ transactions. The incorrect host name that is registered can cause transaction recovery issues.

Workaround:

To fix this, use Registry Editor (regedit) on the remote Windows system and look under the registry key:

HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0

The string-type values for the following key must be set correctly:

Name:            *Host*
Type:            REG_SZ
Data: << the name of the host-machine e.g. myhost@example.com <mailto:myhost@example.com>>> 

This key must contain the host name of the system on which the registry key exists.

This issue is tracked with Oracle bug 9483665.

6.2.21 Oracle Clusterware Upgrade Error

When you use Oracle Universal Installer to upgrade Oracle Clusterware 11g Release 1 (11.1.0.6), or Oracle Clusterware 11g Release 1 (11.1.0.7) to Oracle Clusterware 11g Release 2 (11.2), you may receive the following error:

[INS-40406] There is no previous version of grid software detected on the system.

Workaround:

If you are performing the upgrade on a remote node, a node other than the one on which Oracle Clusterware 11g Release 1 (11.1.0.6) was installed, then the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc may not have been set on that node.

Ensure that you manually set the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc to the location of the Oracle Inventory directory of the node on which you are performing the upgrade installation.

This issue is tracked with Oracle bug 9210861.

6.2.22 Issues Associated with Oracle SID Containing Underscore Character

Oracle Database 11.2 supports Oracle SID having an underscore but it is not compatible with the earlier client releases with a bequeath connection. The issue affects any Oracle Database Client 11.1.0.7 or earlier client, when connecting to the database over a bequeath connection. This includes all OCI clients and Oracle Enterprise Manager agents that monitor the database.

Workaround:

Use only alphanumeric characters including hyphens for Oracle SID or connect to the database using TCP connection.

This issue is tracked with Oracle bug 7611024.

6.2.23 Oracle ARP Protocol Driver Not Installed During Oracle Grid Infrastruction Installation

During Oracle Grid Infrastructure installation, the Oracle ARP protocol driver is not installed by default. You can choose to install the oraarp driver after completing the database installation.

Workaround:

After the Oracle Grid Infrastructure installation is complete, perform the following steps to install the Oracle ARP protocol driver:

1. Start the Oracle Grid Infrastructure stack, if it is not up.

   Use crsctl check crs command to check the status of Oracle Restart or use crsctl check cluster command to check the status of Oracle Grid Infrastructure for a cluster.

   Use crsctl start crs command to start Oracle Restart or crsctl start cluster command to start the Oracle Grid Infrastructure stack.

2. On each node in the cluster, run the following command to start the oraarp driver installation:

   %systemroot%\system32\netcfg.exe -l Grid_home\bin\oraarpdrv.inf -c p 
           -i orcl_ndisprot 
   

3. If "Oracle America, Inc." is not in list of the trusted publishers of the node, then the following message appears:

   Would you like to install this device software? 
   Name: Oracle Network Protocol 
   Publisher: Oracle America, Inc. 
   

   Select Install to install the driver.

4. Run the following command on all the nodes in the cluster after installing the oraarp driver:

   sc config oraarpdrv start= auto
   

5. Stop the Oracle GNS service. To determine if GNS is configured, run the following command:

   Grid_home\bin\srvctl config gns
   

   To stop the GNS service, run the following command:

   Grid_home\bin\srvctl stop gns 
   

6. Stop the SCAN resource using the following command:

   Grid_home\bin\srvctl stop scan -f
   

7. Stop all node applications running on the nodes in the cluster. To determine the node names for your cluster, run the following command:

   Grid_home\bin\olsnodes -l
   

   To stop the node application resources, use the following command where node_list is a comma-delimited list of node names:

   Grid_home\bin\srvctl stop nodeapps -n node_list -f
   

8. Restart the node applications on all nodes in the cluster using the following command where node_list is a comma-delimited list of node names:

   Grid_home\bin\srvctl start nodeapps -n node_list
    
   

9. Restart the SCAN resource using the following command:

   Grid_home\bin\srvctl start scan        
   

If necessary, here are the steps to uninstall Oracle ARP protocol driver:

1. Stop the Oracle ARP protocol driver:

   net.exe stop oraarpdrv
   

2. Open the Windows Networks Connections management interface. Right-click Properties on a public network interface, or from the Start menu, select Control Panel, then Network and Internet, and then Network Connections.

3. Select Oracle ARP Protocol Driver and click Uninstall.

This issue is tracked with Oracle bug 13087098.

6.2.24 Services for an Oracle RAC 10.2 Database Instance Fail to Start

If you run an Oracle Database release 10.2 or 11.1 on Oracle Grid Infrastructure release 11.2, after a node is restarted, the database instance resource running on that node may erroneously report an online state when it is actually offline. Any service associated with the instance fails to start.

Workaround:

Manually issue a start against the instance by running the following commands on the node that was restarted:

srvctl stop instance -d db_unique_name -i instance_name
srvctl start instance -d db_unique_name -i instance_name

This issue is tracked with Oracle bug 10224187.

6.2.25 Oracle ASM Disk Group Does Not Mount on Remote Node After a Database Upgrade on Windows x64

After upgrading Oracle Grid Infrastructure to release 11.2.0.3, the Oracle Automatic Storage Management Configuration Assistant may attempt to start the disk group on a remote node before the disk group is created. As a result, the disk group fails to start.

Workaround:

Manually start the disk group.

This issue is tracked with Oracle bug 13099648.

6.2.26 Oracle Grid Infrastructure Configuration Error on Windows x64

Oracle Grid Infrastructure wizard fails to compare the registered Oracle homes across nodes in case sensitive manner.

Workaround:

Ensure that the Oracle home and Oracle base location specified during a software-only installation across nodes match in a case-sensitive manner.

This issue is tracked with Oracle bug 13242451.

6.2.27 Oracle Universal Installer Does Not Detect Existing Oracle Grid Infrastructure During an Upgrade on Windows x64

If you perform a software-only installation of Oracle Grid Infrastructure release 11.2.0.3 on a system that is currently running 11.2.0.1 release, and then you start the software configuration wizard to upgrade Oracle Grid Infrastructure to 11.2.0.3 release, you may get the following error message:

[INS-40418] The installer has detected that Oracle Clusterware is not running on local node. Oracle recommends that you upgrade Oracle Clusterware in  rolling manner.

Workaround:

Run cluvfy comp clu from the 11.2.0.3 home, before starting the configuration wizard.

This issue is tracked with Oracle bug 13321999.

6.2.28 ocfs.sys Driver Upgrade Issue on Windows x64

When you upgrade Oracle Grid Infrastructure to release 11.2.0.3, the ocfs.sys driver is not upgraded.

Workaround:

Perform the following steps after installing Oracle Grid Infrastructure release 11.2.0.3:

1. Stop all services using OCFS volumes.

2. Run the following command:

   ocfsinstall /r
   

3. Run the following command:

   ocfsinstall /y
   

4. Restart the node for the new OCFS driver to take effect.

This issue is tracked with Oracle bug 13247694.

6.3.1 Issues While Installing .NET Related Products through Mapped Drives on Windows x64

When installing .NET related products such as Oracle Data Provider for .NET, Oracle Providers for ASP.NET, and so on, to an ORACLE_HOME and ORACLE_BASE that are located on a mapped drive on either 32-bit or 64-bit operating system, you may encounter a System.IO.FileLoadException exception from OraProvCfg.exe.

Workaround:

Install .NET Framework 2.0 SP2 or later on the system where you run the installation.

This issue is tracked with Oracle bug 10225338.

6.3.2 Issues While Specifying the Software Updates Download Location on Windows x64

Oracle Universal Installer does not support spaces and double quotation marks in the path of the download location when the Software Updates option is chosen.

This issue is tracked with Oracle bug 10194635.

6.3.3 SQL Developer Shortcut Appears After Oracle Database Client Upgrade on Windows x64

After upgrading the Oracle Database Client release 11.2.0.1 to release 11.2.0.3, the SQL Developer shortcut appears.

Workaround:

Do not use this SQL Developer shortcut as it is not supported.

This issue is tracked with Oracle bug 10078387.

6.3.4 Inventory Issue with addNode.bat Script

After running the addNode.bat script, the inventory on the new nodes does not reflect the newly added home.

Workaround:

On the node being added by the addNode.bat script, do the following:

CRS or Grid Home

cd Grid_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -cfs -local 

Database or Oracle RAC Home

cd RAC_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -cfs -local 

This issue is tracked with Oracle bug 9541191.

6.3.5 Instance Terminates Following ORA-29770 Error on Windows x64

The default registry key setting for remote memory copy timeout (30 seconds) is too long. This results in an ORA-29770 error when the shared server feature is enabled and LMS processes are stuck in the memory copy code path with TCP library on Windows platform.

Workaround:

Enable the remote copy routine to time out in 5 seconds instead of the default 30 seconds and return the control back to LMS. Update the following registry entry:

\HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OSD9i\IPC
    RemoteCopyTimeout    => 5
    RemoteCopyTimeoutRet => 1

This issue is tracked with Oracle bug 13055779.

6.3.6 Shortcuts for Oracle Software Not Removed from Start Menu When Removing a Node

Start menu shortcut entries are not removed when a node is deleted from a shared Oracle RAC home.

Workaround:

Manually remove the Start menu shortcut entries:

1. From the Start menu, select All Programs, then right-click Oracle-HOME_NAME.

2. Select the delete option and confirm the deletion.

This issue is tracked with Oracle bug 10074059.

6.3.7 PATH Variable Not Updated on Recently Deleted Node that Uses a Shared Oracle Home

If your Oracle RAC uses a shared Oracle home and you delete a node, then the PATH environment variable on the node being deleted is not updated to remove the location of the Oracle home.

Workaround:

Update the PATH variable manually on deleted node using the following steps:

1. From the Start menu, select Control Panel, then open System.

2. Select Advanced and click Environment Variables.

3. Select PATH variable from System Variables and double-click to get variable value.

4. Remove Oracle RAC home reference from variable value and click OK.

5. Click OK on Environment Variables dialog and then click Apply on System Properties dialog.

This issue is tracked with Oracle bug 10074038.

6.3.8 Issues While Deleting Oracle RAC on a Remote Node

When using Oracle Database Configuration Assistant to delete an Oracle RAC instance on a remote node, you may encounter the following error:

EMCA failed due to the following error - Error updating EM configuration for node node_name

Workaround:

Use Process Monitor, PMON, to identify and stop ohFilesPerm.pl perl process. Then remove the Oracle Enterprise Manager Database Control state directory:

ORACLE_HOME/nodename_db_unique_name

This issue is tracked with Oracle bug 10073887.

6.3.9 Using Oracle ACFS for the Database Home and LOB Files

If you place the Oracle Database home on Oracle ACFS, and the database uses LOB files, which are also stored on Oracle ACFS, then the Oracle ACFS home must be of the format, DRIVE_LETTER:\directory. For example:

g:\dir
g:\dir1\dir2

Otherwise, LOB related errors are shown when accessing the LOB (as in the sample schema for a basic installation). For example:

ERROR at line 1:
ORA-22288: file or LOB operation FILEOPEN failed
ORA-06512: at "SYS.DBMS_LOB", line 744
ORA-06512: at "XDB.DBMS_XDB", line 301
ORA-06512: at line 4

See "File Systems Mounted with Drive Letters on Oracle ADVM Volumes Are Not Visible in Windows Explorer" for related information. If you are not using a drive letter for the Oracle ACFS home, then several LOB related error messages appear during sample schema install due to this issue.

Workaround:

Click Ignore these errors to proceed, or click Abort to stop the database creation.

This issue is tracked with Oracle bug 9527303.

6.3.10 Download Microsoft Handle Utility When Using Oracle ACFS

The Microsoft handle utility is required for proper operation of Oracle ACFS High Availability resources managed by Oracle Clusterware. Oracle Clusterware resources may not shut down properly if this command is not included.

Workaround:

Download the Microsoft Handle utility from:

http://technet.microsoft.com/en-us/sysinternals/bb896655.aspx

After downloading, place the Handle utility in the BIN directory of your Grid home.

This issue is tracked with Oracle bug 9349433.

6.3.11 Oracle Grid Infrastructure Listeners Issue

After installing Oracle Grid Infrastructure, the default listener does not listen on public addresses. As a result, even though the listener is available, the listener does not service the database or Oracle ASM instances.

Workaround 1:

If you are not using GNS with DHCP, then perform the following steps:

1. Add the local IP address and host name information to %Systemroot%\System32\Drivers\Etc\Hosts.

   For example, IP address host name such as 192.0.2.254 dbhost.

2. Restart the Oracle Grid Infrastructure listener.

This issue is tracked with Oracle bug 9382523.

Workaround 2:

If you have configured GNS and use DHCP to assign the addresses for the cluster nodes, then configure DHCP so that the IP address assigned to the host name for each cluster node (the public IP address) has a smaller numeric value than the IP addresses assigned for the Virtual IP (VIP) address, GNS VIP address, or SCAN VIP address for a node.

For example, you could use IP addresses in the range of xxx.xxx.xxx.0 – xxx.xxx.xxx.099 for only host IP addresses, and all other IP address assignments must use a value of xxx.xxx.xxx.100 or higher.

This issue is tracked with Oracle bug 9441288.

6.3.12 ODP.NET Help Not Installed on 64-Bit Windows

If you install 32-bit Oracle software on a 64-bit Windows computer, then help for ODP.NET is not installed.

Workaround:

Use the .NET Help Setup Wizard to manually install the dynamic help. To start the wizard on an English locale, run the DataProviderHelp_en.msi file from the ORACLE_BASE\ORACLE_HOME\ODP.NET\help directory. For the Japanese locale, run the DataProviderHelp_ja.msi file from the same directory.

This issue is tracked with Oracle bug 9524039.

6.3.13 Help File Shortcut Not Working in Windows Server 2008

Starting Oracle Objects for OLE C++ Class Library Help from the Start menu shortcut does not open the help page in Windows Server 2008.

Workaround:

Complete the following steps to resolve this issue:

1. Right-click the shortcut.

2. Select Open With and click Internet Explorer.

3. Select the option to open the help page.

This issue is tracked with Oracle bug 8001859.

6.3.14 Shared Oracle Home Shortcuts Do Not Work on Windows Server 2008

Starting Oracle tools such as Net Configuration Assistant, Database Configuration Assistant, Database Upgrade Assistant, Net Manager, Oracle Universal Installer, and so on, from a shared Oracle home shortcut gives the following error:

Missing shortcut: windows is searching launch.exe

After displaying the error message, the shortcut exits.

Workaround:

Start the tools from a Windows command prompt with Administrator privileges.

This issue is tracked with Oracle bug 7525053.

6.3.15 Installing Online Patches on Windows Systems

The Opatch tool shipped with Oracle Database 11g Release 2 fails while applying online patches on Windows systems.

Workaround:

Opatch tool version 11.2.0.1.9 is available on My Oracle Support as Patch 6880880 for Oracle Database 11g Release 2. Download and install Opatch Tool version 11.2.0.1.9 before applying online patches on Windows systems.

To list the database instances patched by an online patch, run the opatch lsinventory -details command. Ensure that you run the opatch rollback command instead of opatch disableOnlinepatch command to disable and remove an online patch from the final database instance in a node. To remove the online patch that is in a disabled state, run the disableOnlinepatch command. To remove the online patch in a disabled state, use -no_sysmod flag with opatch rollback command, only when opatch lsinv -details does not list any instances for that patch. For example, use opatch rollback -id 1234 -no_sysmod, only when opatch lsinv -details show no instances patched for bug 1234.

This issue is tracked with Oracle bug 9440270.

6.3.16 Oracle ACFS Replication Background Processes Issue on Windows x64

When replicating an Oracle ACFS file system, the acfsutil bg stop command does not stop the replication background processes on the sites hosting the primary and standby file systems.

Workaround:

Perform the following steps to dismount the file system successfully or bring the Oracle Grid Infrastructure stack down on a node:

Primary Site

1. Run the following commands:

   crsctl stop res repl_transport_resource -n node
   crsctl relocate res repl_monitor_resource
   crsctl relocate res repl_init_resource
   

2. Stop any acfsrepl_transport daemon running on this node.

3. Manually unmount the file system on this node

Standby Site

1. Run the following commands:

   crsctl stop res repl_transport_resource -n node
   crsctl relocate res repl_monitor_resource
   crsctl relocate res repl_preapply_resource
   crsctl relocate res repl_apply_resource
   

2. Stop any acfsrepl_transport daemon running on this node.

3. Manually unmount the file system on this node

Perform the following steps if any one of these situations apply:

• There is only one node with the file system mounted in the Oracle RAC cluster.

• All the nodes in the cluster must unmount the file system or bring the Oracle Grid Infrastructure stack down.

Primary Site

1. Run the following commands:

   crsctl stop res repl_transport_resource -n node
   crsctl stop res repl_monitor_resource
   crsctl stop res repl_init_resource
   

2. Stop any acfsrepl_* daemon running on any node.

3. Manually unmount the file system on all the nodes before the Oracle Grid Infrastructure stack is down.

Standby Site

1. Run the following commands:

   crsctl stop res repl_transport_resource -n node
   crsctl stop res repl_monitor_resource
   crsctl stop res repl_preapply_resource
   crsctl stop res repl_apply_resource
   

2. Stop any acfsrepl_* daemon running on any node.

3. Manually unmount the file system on all the nodes before the Oracle Grid Infrastructure stack is down.

This issue is tracked with Oracle bug 13247273.

7.2.1 Oracle Configuration Manager Service Not Removed on Remote Node When Performing Cluster Deinstallation on Windows x64

When removing Oracle Configuration Manager from an Oracle home, in some cases, the service is stopped on remote nodes but not removed.

Workaround:

Remove the service using the sc command:

sc remove OracleORACLE_HOME_IDENTIFIERConfigurationManager

This issue is tracked with Oracle bug 10074102.

7.2.2 Service Not Removed When Deleting a Node on Shared Oracle RAC

When deleting a node on shared Oracle RAC home, OraClrAgnt service is stopped but not removed.

Workaround:

Manually remove the service using the sc command:

sc remove OracleORACLE_HOME_NAMEClrAgent

This issue is tracked with Oracle bug 10074070.

7.2.3 Issues While Installing Oracle Database with Oracle ASM on Windows x64

If you install Oracle Database release 11.1 or earlier and during installation choose to create a database with storage configured on Oracle ASM release 11.2.0.4, then the installation fails while the configuration tools are running, specifically Database Configuration Assistant.

Workaround:

Install the Oracle Database release 11.1 or earlier software and choose the software only installation option. After the installation completes, create the single-instance database by launching Database Configuration Assistant from the newly installed Oracle home. For example, run the following command from a command prompt:

ORACLE_BASE\ORACLE_HOME\bin\dbca.bat

This issue is tracked with Oracle bug 10149589.

7.2.4 Deinstallation Tool Fails to Remove Shared Oracle RAC Home on Windows x64

When using the Deinstallation tool, deinstall.bat, to remove a shared Oracle RAC home, you get the error message, Failed to delete directory.., for the Oracle home being removed.

Workaround:

Use Process Monitor, PMON, to locate and stop the processes running out of the Oracle home being removed. Then remove the directory.

This issue is tracked with Oracle bug 10145215.

7.2.5 Deinstallation Tool Issues with Common Oracle Base Location on Windows x64

If you install a 32-bit Oracle Database and a 64-bit Oracle Database in the same ORACLE_BASE, then it may lead to unexpected results when using the deinstallation tool to remove one of the databases. The deinstallation tool removes all the Oracle home under Oracle base if these Oracle homes do not use the same central inventory.

Workaround:

Avoid using multiple central inventories. Do not use the same ORACLE_BASE for 32-bit and 64-bit database installations.

This issue is tracked with Oracle bugs 12808859 and 12864734.

7.2.6 Issues While Installing Oracle RAC with Oracle ASM on Windows x64

If you install Oracle RAC release 11.2.0.1 with Oracle ASM release 11.2.0.4, then the following error may occur:

OUI-35078: Error in updating env on nodes 'node_name'.

Workaround:

Perform the following steps:

1. Run the following command from the 11.2.0.4 Grid home:

   srvctl stop cvu
   

2. Stop the Oracle Enterprise Manager Database Control using the following command:

   emctl stop agent
   

3. Stop the OracleRemExecService from all the nodes.

4. Then, install Oracle RAC 11.2.0.1 with Oracle ASM release 11.2.0.4.

This issue is tracked with Oracle bug 12720534.

7.2.7 Oracle Grid Infrastructure Release 11.2.0.4 Upgrade Issues with Oracle RAC Release 11.2.0.1

When you upgrade Oracle Grid Infrastructure to release 11.2.0.4 and start installing Oracle RAC release 11.2.0.1, then the Oracle Database Configuration Assistant shows the following error message towards the end of the database creation:

java.exe - Entry Point Not Found
 
The procedure entry point clsunsinit could not be located in the dynamic link library orahasgen11.dll.
---------------------------
OK  
---------------------------

Workaround:

Click OK and proceed with the database creation process. Log into the Oracle Database Enterprise Manager Control and navigate to the Cluster home page. Click Monitoring Configuration and update the eONS port for cluster target property.

This issue is tracked with Oracle bug 13074121.

7.2.8 Oracle Universal Installer Does Not Detect Existing Oracle Grid Infrastructure During an Upgrade on Windows x64

If you perform a software-only installation of Oracle Grid Infrastructure release 11.2.0.4 on a system that is currently running 11.2.0.1 release, and then you start the software configuration wizard to upgrade Oracle Grid Infrastructure to 11.2.0.4 release, you may get the following error message:

[INS-40418] The installer has detected that Oracle Clusterware is not running on local node. Oracle recommends that you upgrade Oracle Clusterware in  rolling manner.

Workaround:

Run cluvfy comp clu from the 11.2.0.4 home, before starting the configuration wizard.

This issue is tracked with Oracle bug 13321999.

7.2.9 Deinstallation Tool Fails to Remove Registry Entries

When deinstalling Oracle Grid Infrastructure, some registry entries are not removed.

Workaround:

Manually remove the following registry entries:

HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\ODP.NET
HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\_DosLinks
HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Oracle11

This issue is tracked with Oracle bug 16982469.

Manually remove the registry entry, OracleDBConsoleSID:

HKLM\System\CurrentControlSet\Services\EventLog\Application
 

This issue is tracked with Oracle bug 9523875.

7.2.10 Issues Associated with Oracle SID Containing Underscore Character

Oracle Database 11.2 supports Oracle SID having an underscore but it is not compatible with the earlier client releases with a bequeath connection. The issue affects any Oracle Database Client 11.1.0.7 or earlier client, when connecting to the database over a bequeath connection. This includes all OCI clients and Oracle Enterprise Manager agents that monitor the database.

Workaround:

Use only alphanumeric characters including hyphens for Oracle SID or connect to the database using TCP connection.

This issue is tracked with Oracle bug 7611024.

7.2.11 Oracle ARP Protocol Driver Not Installed During Oracle Grid Infrastruction Installation

During Oracle Grid Infrastructure installation, the Oracle ARP protocol driver is not installed by default. You can choose to install the oraarp driver after completing the database installation.

Workaround:

After the Oracle Grid Infrastructure installation is complete, perform the following steps to install the Oracle ARP protocol driver:

1. Start the Oracle Grid Infrastructure stack, if it is not up.

   Use crsctl check crs command to check the status of Oracle Restart or use crsctl check cluster command to check the status of Oracle Grid Infrastructure for a cluster.

   Use crsctl start crs command to start Oracle Restart or crsctl start cluster command to start the Oracle Grid Infrastructure stack.

2. On each node in the cluster, run the following command to start the oraarp driver installation:

   %systemroot%\system32\netcfg.exe -l Grid_home\bin\oraarpdrv.inf -c p 
           -i orcl_ndisprot 
   

3. If "Oracle America, Inc." is not in list of the trusted publishers of the node, then the following message appears:

   Would you like to install this device software? 
   Name: Oracle Network Protocol 
   Publisher: Oracle America, Inc. 
   

   Select Install to install the driver.

4. Run the following command on all the nodes in the cluster after installing the oraarp driver:

   sc config oraarpdrv start= auto
   

5. Stop the Oracle GNS service. To determine if GNS is configured, run the following command:

   Grid_home\bin\srvctl config gns
   

   To stop the GNS service, run the following command:

   Grid_home\bin\srvctl stop gns 
   

6. Stop the SCAN resource using the following command:

   Grid_home\bin\srvctl stop scan -f
   

7. Stop all node applications running on the nodes in the cluster. To determine the node names for your cluster, run the following command:

   Grid_home\bin\olsnodes -l
   

   To stop the node application resources, use the following command where node_list is a comma-delimited list of node names:

   Grid_home\bin\srvctl stop nodeapps -n node_list -f
   

8. Restart the node applications on all nodes in the cluster using the following command where node_list is a comma-delimited list of node names:

   Grid_home\bin\srvctl start nodeapps -n node_list
    
   

9. Restart the SCAN resource using the following command:

   Grid_home\bin\srvctl start scan        
   

If necessary, here are the steps to uninstall Oracle ARP protocol driver:

1. Stop the Oracle ARP protocol driver:

   net.exe stop oraarpdrv
   

2. Open the Windows Networks Connections management interface. Right-click Properties on a public network interface, or from the Start menu, select Control Panel, then Network and Internet, and then Network Connections.

3. Select Oracle ARP Protocol Driver and click Uninstall.

This issue is tracked with Oracle bug 13087098.

7.2.12 Oracle MTS Recovery Service Issue When Installing Oracle RAC

In an Oracle RAC installation, Oracle MTS Recovery Service on the remote nodes does not have the correct host name registered for string variable HOST under registry key HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0. This variable contains the name of the host from where the installation was performed instead of the name of the local host. This registry variable is used by COM+ applications to pass information to Oracle databases so that they can connect back to the Windows computer if the database must resolve in-doubt COM+ transactions. The incorrect host name that is registered can cause transaction recovery issues.

Workaround:

To fix this, use Registry Editor (regedit) on the remote Windows system and look under the registry key:

HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OracleMTSRecoveryService\Protid_0

The string-type values for the following key must be set correctly:

Name:            *Host*
Type:            REG_SZ
Data: << the name of the host-machine e.g. myhost@example.com <mailto:myhost@example.com>>> 

This key must contain the host name of the system on which the registry key exists.

This issue is tracked with Oracle bug 9483665.

7.2.13 Oracle Clusterware Upgrade Error

When you use Oracle Universal Installer to upgrade Oracle Clusterware 11g Release 1 (11.1.0.6), or Oracle Clusterware 11g Release 1 (11.1.0.7) to Oracle Clusterware 11g Release 2 (11.2), you may receive the following error:

[INS-40406] There is no previous version of grid software detected on the system.

Workaround:

If you are performing the upgrade on a remote node, a node other than the one on which Oracle Clusterware 11g Release 1 (11.1.0.6) was installed, then the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc may not have been set on that node.

Ensure that you manually set the registry key HKEY_LOCAL_MACHINE\SOFTWARE\Oracle\inst_loc to the location of the Oracle Inventory directory of the node on which you are performing the upgrade installation.

This issue is tracked with Oracle bug 9210861.

7.2.14 Oracle ASM Disk Group Does Not Mount on Remote Node After a Database Upgrade on Windows x64

After upgrading Oracle Grid Infrastructure to release 11.2.0.4, the Oracle Automatic Storage Management Configuration Assistant may attempt to start the disk group on a remote node before the disk group is created. As a result, the disk group fails to start.

Workaround:

Manually start the disk group.

This issue is tracked with Oracle bug 13099648.

7.2.15 PRKN-1019 Error During Oracle Grid Infrastructure Upgrade to Release 11.2.0.4 on Windows x64

While upgrading Oracle Grid Infrastructure from release 11.2.0.1 to release 11.2.0.4, the following errors are reported:

OUI-35078: Error in updating env on nodes 'node_name'.
 
[PRKN-1019 : Failed to create directory
""%TMPDIR%\oraremservice\" on node
"node_name", Error: "0|The network path was not found. " .]

However, the specified directory exists on the node being upgraded.

Workaround:

Perform the following steps before restarting the Oracle Grid Infrastructure upgrade:

1. Stop the OracleRemExecService on all the nodes in a cluster.

2. Stop the dbconsole process and EMagent on all the nodes in a cluster by running the following command on each node:

   emctl stop dbconsole
   

This issue is tracked with Oracle bug 13052343.

7.2.16 Oracle ACFS File Systems Must be Manually Dismounted Before Upgrade, Deinstallation, or Direct Shutdown of Oracle Clusterware or Oracle ASM

When doing an upgrade, deinstallation, or direct shutdown of Oracle Clusterware or Oracle ASM, to ensure that Oracle Clusterware (or Oracle ASM) comes down cleanly, you must dismount all Oracle ACFS file systems manually before Oracle Cluserware shutdown. Use the handle and wmic commands to identify processes which are active on the Oracle ACFS file systems. To ensure that these processes are no longer active, dismount all Oracle ACFS file systems and issue Oracle Clusterware shutdown. Otherwise, errors may be issued during Oracle Clusterware shutdown relating to activity on Oracle ACFS file systems which will stop the successful shutdown of Oracle Clusterware.

This issue is tracked with Oracle bug 8594128.

7.2.17 Duplicate SID Not Detected

While installing Oracle RAC, Oracle Universal Installer does not detect duplicate SID values across nodes.

Workaround:

Ensure that you specify a unique SID across nodes.

This issue is tracked with Oracle bug 16862658.

7.2.18 Deinstallation Does Not Remove Oracle Enterprise Manager Configuration Assistant (EMCA) Properties File

When deinstalling Oracle Database or Oracle RAC, ORACLE_HOME\sysman\config\emca.properties file may not get deleted occasionally.

Workaround:

Manually delete the emca.properties file from ORACLE_HOME\sysman\config\, and then delete the ORACLE_HOME directory.

This issue is tracked with Oracle bug 16968572.

7.2.19 Microsoft Visual C++ 2005 Redistributable Package Required for the Deinstallation Tool

If the system administrator removes the Microsoft Visual C++ 2005 Package after installing Oracle Database, then running the deinstallation tool shows the following error message:

OraOLEDBPus11.dll: The specified module could not be found

Workaround:

Reinstall the Microsoft Visual C++ 2005 Redistributable Package.

This issue is tracked with 12797310.

7.2.20 Remote Agents Must Be Manually Reloaded After Cluster Ready Services Upgrade and Oracle ASM Upgrade

When attempting to upgrade Cluster Ready Services or Oracle ASM to 11.2 release using the Automatic Storage Management Configuration Assistant, the upgrade succeeds, but the configuration changes made to the Agents on a remote host fail to reload. As a result, Oracle Enterprise Manager Grid Control and Database Control cannot monitor the Oracle ASM and Cluster Ready Services targets.

Workaround:

Manually reload the remote agents after the upgrade using the following command:

emctl reload agent

This issue is tracked with Oracle bug 9398380.

7.2.21 Deinstallation Tool Does Not Remove Oracle Home

When using the Deinstallation Tool to remove an Oracle home that has Oracle Enterprise Manager Database Control configured and that was installed on Oracle ACFS, the ORACLE_BASE\ORACLE_HOME\host_sid\sysman directory and its sub directories cannot be deleted, preventing the Oracle home from being removed.

Workaround:

Before performing these operations, run the following command as the Administrator user:

cacls ORACLE_BASE\ORACLE_HOME\HOSTNAME_DBNAME\sysman\emd\upload\hsperfdata_SYSTEM /E /P %USERDOMAIN%\%USERNAME%:F 

This issue is tracked with Oracle bug 9444816.

7.2.22 Deinstallation Tool Does Not Remove invDetails.properties File

The Deinstallation Tool fails to remove the invDetails.properties file when deinstalling Oracle RAC home.

Workaround:

Manually remove the dbhome\inventory directory.

This issue is tracked with Oracle bug 9543109.

7.2.23 Using Deinstallation Tool for a Single-Instance Oracle Database Returns Errors

Removing Oracle home fails with file in use errors as the services and processes running from the Oracle home are not removed completely by Microsoft's Service Control Manager before the Oracle home removal.

Workaround:

Remove the Oracle home manually after the deinstall tool exits.

This issue is tracked with Oracle bug 10035858.

7.2.24 Auto-Update Fails on a Shared Oracle RAC Home

With shared Oracle RAC homes, if you use the installer auto-update feature to download and apply the latest patches from My Oracle Support, the installer fails to apply the patches with a Java runtime Exception error.

Workaround:

Do not use the auto-update option during installation and apply the required patches using Opatch after the installation. Alternatively, if you use the auto-update option, after you have completed installation, manually download the patches which failed to be applied, from My Oracle Support, and then run the following command to apply the patches:

ORACLE_HOME\OPatch\opatch napply -local

This issue is tracked with Oracle bug 17501170.

7.2.25 Listener, SCAN Listener, and SCAN VIP Resources Fail to Start After Database Upgrade

After upgrading Oracle Grid Infrastructure or Oracle RAC from release 10.2.0.4 to 11.2.0.4, the listener, SCAN listener, and SCAN VIP resources fail to start when you restart the node.

Workaround:

Manually install the Oracle ARP protocol driver and run the following command to start the oraarp driver automatically when restarting the node:

sc config oraarpdrv start= auto

This issue is tracked with Oracle bug 17084320.

7.2.26 No Available Nodes Listed During Oracle Grid Infrastructure Upgrades

During Oracle Grid Infrastructure upgrades, or during installations on servers where previous releases of Oracle Grid Infrastructure were installed, the Cluster Node Information window may indicate that there are no available cluster node candidates. The available nodes are not listed on systems with the registry key \HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OSD9i\IPC. The registry key was a part of the configuration requirements for the previous releases for Oracle cluster software, but it is not used with Oracle Grid Infrastructure.

Workaround:

On each cluster node, remove the following registry entries to use the default values:

\HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\OSD9i\IPC
     RemoteCopyTimeout    => 5
     RemoteCopyTimeoutRet => 1

After modifying the registry, restart all cluster nodes, and continue with the installation.

This issue is tracked with Oracle bug 18504336.

7.2.27 INS-13001 Error on Windows Server 2012 R2

Installing Oracle Database using the command, setup.exe -jreLoc path-to-jdk7\jre results in the following error:

INS-13001: Environment does not meet minimum requirements.

Workaround:

Skip the preinstallation requirements check using the following command:

setup.exe -jreLoc path-to-jdk7\jre -ignorePrereq

This issue is tracked with Oracle bug 18435100.

7.2.28 Oracle ACFS Support on Windows Server 2012 R2

When creating an Oracle home on Oracle ACFS on Windows Server 2012 R2, the clusterware resources associated with the home may be in an ITERMEDIATE (versus ONLINE) state on selected nodes in the cluster due to an acfsutil error, as shown in the following example:

ora.acfstest.dbhome.acfs
        ONLINE  ONLINE         node_1 mounted on L:\
        ONLINE  ONLINE         node_2 mounted on L:\
        ONLINE  INTERMEDIATE   node_3 acfsutil error

Workaround:

This issue will be fixed in an upcoming patch bundle.

This issue is tracked with Oracle bug 19478594.

7.3.1 Issues When Using addnode.bat

When using the addnode script and specifying a response file, you may see the following error:

ERROR: Value for CLUSTER_NEW_NODES not specified.

Workaround:

Do not use a response file.

This issue is tracked with Oracle bug 17167249.

If you use fully-qualified domain names for the node as the value you specified in addnode.bat for the CLUSTER_NEW_NODES parameter, then the Database Configuration Assistant cannot retrieve the correct information for Oracle RAC instance management. This occurs if you add a node to Oracle Grid Infrastructure for a Cluster, and if you add Oracle RAC instances to cluster nodes. For example:

 addNode.bat "CLUSTER_NEW_NODES={rac2.example.com}"
 "CLUSTER_NEW_VIRTUAL_HOSTNAMES={rac2-vip.example.com}"


addNode.bat "CLUSTER_NEW_NODES={rac2.example.com}"

Workaround:

Update the ORACLE_HOME node list with a domain name that is not fully qualified before you run Oracle Database Configuration Assistant. For example:

setup.exe -updateNodeList
ORACLE_HOME="c:\app\oracle\product\11.2.0\dbhome_1"
 "CLUSTER_NODES={rac1,rac2}" 

This issue is tracked with Oracle bug 17167776.

7.3.2 Help File Shortcut Not Working in Windows Server 2008

Starting Oracle Objects for OLE C++ Class Library Help from the Start menu shortcut does not open the help page in Windows Server 2008.

Workaround:

Complete the following steps to resolve this issue:

1. Right-click the shortcut.

2. Select Open With and click Internet Explorer.

3. Select the option to open the help page.

This issue is tracked with Oracle bug 8001859.

7.3.3 Inventory Issue with addNode.bat Script

After running the addNode.bat script, the inventory on the new nodes does not reflect the newly added home.

Workaround:

On the node being added by the addNode.bat script, do the following:

CRS or Grid Home

cd Grid_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" CRS=TRUE  -cfs -local 

Database or Oracle RAC Home

cd RAC_home\oui\bin

If you have a non-shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -local

If you have a shared home, then run the following command:

attachHome.bat "CLUSTER_NODES={comma separated node list}" -cfs -local 

This issue is tracked with Oracle bug 9541191.

7.3.4 Using Oracle ACFS for the Database Home and LOB Files

If you place the Oracle Database home on Oracle ACFS, and the database uses LOB files, which are also stored on Oracle ACFS, then the Oracle ACFS home must be of the format, DRIVE_LETTER:\directory. For example:

g:\dir
g:\dir1\dir2

Otherwise, LOB related errors are shown when accessing the LOB (as in the sample schema for a basic installation). For example:

ERROR at line 1:
ORA-22288: file or LOB operation FILEOPEN failed
ORA-06512: at "SYS.DBMS_LOB", line 744
ORA-06512: at "XDB.DBMS_XDB", line 301
ORA-06512: at line 4

See "File Systems Mounted with Drive Letters on Oracle ADVM Volumes Are Not Visible in Windows Explorer" for related information. If you are not using a drive letter for the Oracle ACFS home, then several LOB related error messages appear during sample schema install due to this issue.

Workaround:

Click Ignore these errors to proceed, or click Abort to stop the database creation.

This issue is tracked with Oracle bug 9527303.

7.3.5 PATH Variable Not Updated on Recently Deleted Node that Uses a Shared Oracle Home

If your Oracle RAC uses a shared Oracle home and you delete a node, then the PATH environment variable on the node being deleted is not updated to remove the location of the Oracle home.

Workaround:

Update the PATH variable manually on deleted node using the following steps:

1. From the Start menu, select Control Panel, then open System.

2. Select Advanced and click Environment Variables.

3. Select PATH variable from System Variables and double-click to get variable value.

4. Remove Oracle RAC home reference from variable value and click OK.

5. Click OK on Environment Variables dialog and then click Apply on System Properties dialog.

This issue is tracked with Oracle bug 10074038.

7.3.6 Shortcuts for Oracle Software Not Removed from Start Menu When Removing a Node

Start menu shortcut entries are not removed when a node is deleted from a shared Oracle RAC home.

Workaround:

Manually remove the Start menu shortcut entries:

1. From the Start menu, select All Programs, then right-click Oracle-HOME_NAME.

2. Select the delete option and confirm the deletion.

This issue is tracked with Oracle bug 10074059.

7.3.7 Issues While Specifying the Software Updates Download Location on Windows x64

Oracle Universal Installer does not support spaces and double quotation marks in the path of the download location when the Software Updates option is chosen.

This issue is tracked with Oracle bug 10194635.

7.3.8 Issues While Installing .NET Related Products through Mapped Drives on Windows x64

When installing .NET related products such as Oracle Data Provider for .NET, Oracle Providers for ASP.NET, and so on, to an ORACLE_HOME and ORACLE_BASE that are located on a mapped drive on either 32-bit or 64-bit operating system, you may encounter a System.IO.FileLoadException exception from OraProvCfg.exe.

Workaround:

Install .NET Framework 2.0 SP2 or later on the system where you run the installation.

This issue is tracked with Oracle bug 10225338.

7.3.9 Issues While Deleting Oracle RAC on a Remote Node

When using Oracle Database Configuration Assistant to delete an Oracle RAC instance on a remote node, you may encounter the following error:

EMCA failed due to the following error - Error updating EM configuration for node node_name

Workaround:

Use Process Monitor, PMON, to identify and stop ohFilesPerm.pl perl process. Then remove the Oracle Enterprise Manager Database Control state directory:

ORACLE_HOME/nodename_db_unique_name

This issue is tracked with Oracle bug 10073887.

7.3.10 Issue While Running clone.pl

On a Windows 32-bit system, where the msvcr71.dll file is not present in C:\Windows\System32, running the clone.pl script shows the following message:

The application had failed to start because MSVCR71.dll was not found. Reinstalling the application may fix the problem

Workaround:

Set the path of the msvcr71.dll file to include ORACLE_HOME\oui\lib\win32 from a command prompt. Run clone.pl from the same command prompt.

This issue is tracked with Oracle bug 17233850.

7.3.11 ODP.NET Help Not Installed on 64-Bit Windows

If you install 32-bit Oracle software on a 64-bit Windows computer, then help for ODP.NET is not installed.

Workaround:

Use the .NET Help Setup Wizard to manually install the dynamic help. To start the wizard on an English locale, run the DataProviderHelp_en.msi file from the ORACLE_BASE\ORACLE_HOME\ODP.NET\help directory. For the Japanese locale, run the DataProviderHelp_ja.msi file from the same directory.

This issue is tracked with Oracle bug 9524039.

7.3.12 Download Microsoft Handle Utility When Using Oracle ACFS

The Microsoft handle utility is required for proper operation of Oracle ACFS High Availability resources managed by Oracle Clusterware. Oracle Clusterware resources may not shut down properly if this command is not included.

Workaround:

Download the Microsoft Handle utility from:

http://technet.microsoft.com/en-us/sysinternals/bb896655.aspx

After downloading, place the Handle utility in the BIN directory of your Grid home.

This issue is tracked with Oracle bug 9349433.

7.3.13 Shared Oracle Home Shortcuts Do Not Work on Windows Server 2008

Starting Oracle tools such as Net Configuration Assistant, Database Configuration Assistant, Database Upgrade Assistant, Net Manager, Oracle Universal Installer, and so on, from a shared Oracle home shortcut gives the following error:

Missing shortcut: windows is searching launch.exe

After displaying the error message, the shortcut exits.

Workaround:

Start the tools from a Windows command prompt with Administrator privileges.

This issue is tracked with Oracle bug 7525053.

7.3.14 SQL Developer Shortcut Appears After Oracle Database Client Upgrade on Windows x64

After upgrading the Oracle Database Client release 11.2.0.1 to release 11.2.0.4, the SQL Developer shortcut appears.

Workaround:

Do not use this SQL Developer shortcut as it is not supported.

This issue is tracked with Oracle bug 10078387.

7.3.15 Deinstallation Issue with IPv6

On Windows Server 2012, ensure that IPv6 is disabled on all the interfaces, including the interfaces that are not used, otherwise deinstallation fails with the following error:

PRKC-1071: Nodes \"{0}\" did not respond to ping in \"{1}\" seconds, {2}

This issue is tracked with Oracle bug 17263057.

7.3.16 High CPU Usage of a Host Process on Windows Server 2012

On Windows Server 2012, after installing Oracle Grid Infrastructure, you may have a high CPU usage issue for the Wmiprvse.exe process, depending upon the power of the processor.

Workaround:

None

This issue is tracked with Oracle bug 17395707.

7.3.17 Issues Adding Database Service with SRVCTL

When attempting to add Oracle RAC service release 11.1.0.7 on Oracle Grid Infrastructure release 11.2.0.4, the following error occurs:

Exception in thread "main" java.lang.NoSuchFieldError: SCLS_UNEXPECTED_ERROR

Workaround:

Ensure that the current working directory is not the bin directory of 11.2.0.4 Oracle Grid Infrastructure home, GI_HOME\bin, when you execute the srvctl add service command for the 11.1.0.7. Oracle RAC.

This issue is tracked with Oracle bug 17573010.

7.3.18 Database Fails to Start After Private NIC Failure of One Node in a Two Node Cluster

In a two node cluster, when there is a failure of the private NIC on the node that has smallest node number (node A) after the other node B was evicted, then when the private network is reestablished on node A, the other node B joins the cluster but the database instance on the other node is unable to start.

Workaround:

Perform the following steps:

1. Stop the crs stack on both nodes, then start the crs stack on both nodes.

2. Stop the database instance on the node that has the private network failure (node A), then start the database on both nodes. The database instance fails to start on node A this time. After the instance starts on the other node B, then you will be able to start the instance on node A.

This issue is tracked with Oracle bug 19642240.

7.3.19 Cloning Issue on Windows Server 2012 R2

When you clone an Oracle home that has patch 20043415, it may result in Oracle ASM configuration issues with Oracle Grid Infrastructure or with Standard or Enterprise editions of Oracle Database.

Workaround:

Perform the following steps:

1. Unzip the Oracle Grid home or Oracle RAC home on the destination node.

2. Remove the following files:

   ORACLE_HOME\bin\orancrypt11_std.dll.dbl
   ORACLE_HOME\bin\orancrypt11_ee.dll.dbl
   

3. Proceed to clone the Oracle Grid or Oracle RAC home.

This issue is tracked with Oracle bug 19711300.