Release 2 (9.0.2) for UNIX
Part Number A96639-01
 Home |
| Oracle Enterprise Manager Release Notes Release 2 (9.0.2) for UNIX Part Number A96639-01 |
|
April 2002
Part No. A96639-01
This document summarizes the differences between Oracle Enterprise Manager and its documented functionality. Specifically, this document includes the release notes for the Enterprise Manager software components installed with the Oracle9iAS Infrastructure installation type.
The following sections include important information about system requirements that must be addressed before you run the Enterprise Manager software.
In order to run Intelligent Agent Release 9.0.2 on Sun SPARC Solaris 2.6 and 2.7, the following Solaris operating system patches need to be applied:
On Solaris 2.6:
On Solaris 2.7:
Please contact Sun's Technical Support to procure the above mentioned patches.
The JRE is installed by various Oracle products. However, there is only one shared JRE per system. As a result, if after installing Enterprise Manager Release 9.0 you install a previous release of Enterprise Manager (i.e. Release 2.2), the JRE (Java Runtime Environment) installed as part of Release 9.0 will be downgraded automatically; no error messages will appear. This downgrade from JRE Release 1.1.8_12 to 1.1.8_10 will cause problems for Enterprise Manager Release 9.0. Consequently, do no install Enterprise Manager Release 2.2 after installing Enterprise Manager Release 9.0.
Installing a pre-9.0 Oracle database will also cause the same problem of downgrading the JRE. If you run into this situation, you can workaround the problem by re- installing JRE 1.1.8_12 from the CD-ROM you installed Enterprise Manager Release 9.0 (most likely from a Release 9.0 CD-ROM). Re-install the JRE by re-installing Oracle Universal Installer; as the Installer has a dependency on the JRE.
This section describes general issues and their workarounds for Enterprise Manager components installed with the Oracle9iAS Infrastructure.
The following issues and workarounds apply to the Oracle Enterprise Manager Console.
If you select a different window after launching the Console, the login dialog will not display in the foreground. You may need to minimize any active windows or use ALT -> tab and select the Console application to bring the login dialog to the foreground. (Bug 806713)
During discovery of remote nodes and their targets, if the node is available and reachable via the network, but the Intelligent Agent on that node is down, the discovery wizard usually displays the following error:
VNI-4009 : Cannot contact agent on the node. Agent may be down or network communication to the node has failed.
However, in some cases, the following error appears instead:
VNI-4044 : Cannot contact agent. Node may be down, or the network may be down or slow.
If this alternate error occurs, log on to the remote node and check the status of the Intelligent Agent via the agentctl utility:
agenctl status agent
If the Agent is down, start the Agent using the following command and try performing the discovery operation again (Bug 1874173):
agentctl start agent
When you open a modal dialog box in an Oracle Enterprise Manager application (a modal dialog box is one that requires user action before being able to return to the application) and request context-sensitive help on that dialog box, the help is displayed in a modal help topic window. Before you can return to the modal dialog box, you must close the help topic window that is displaying the help topic for the modal dialog box.
For non-English versions of Enterprise Manager, certain inconsistencies may appear in the non-English versions of the online help.
When connecting to a new OMS and before painting the right hand side of the console window (the detail panel), the Enterprise Manager console may stop responding. The workaround is to kill the process and restart. (bug 1702853)
When installing an Oracle9i Agent on a machine running a pre-9i database, you must re-run a version-specific copy of catsnmp.sql which is located in the ORACLE_HOME/sysman/admin directory of the Oracle9i Enterprise Manager client install. For example, if you are running Oracle 8.1.7 on a machine and then install a 9i Agent, you must re-run thecatsnmp_8i.sql script after installing the new Agent. This operation must be performed for each pre-9i database serviced by this Agent.
Do not run the Oracle9i version of the catsnmp.sql script against pre-9i databases.
In non-English deployments of Enterprise Manager on Solaris (and possibly other UNIX based platforms), launching the Netscape browser to show HTML files from an Enterprise Manager application may not work.
For instance, if you select Latest Product Information from the Console Help menu, nothing appears. (Bugs 2188802 and 2191034)
To workaround this problem, perform the following steps.
altbrowser that contains the following:
#!/bin/sh ns=/usr/local/bin/netscape # ---> ns is the absolute path of the actual browser. unset XFILESEARCHPATH # this is the workaround. $ns "$@" "$@" should be used instead of $* # to allow white spaces as the argument
% chmod 755 /usr/local/bin/altbrowser
ClientConfig.properties file, add the following line to the end of the file:
user.browser=/usr/local/bin/altbrowser
The ClientConfig.properties file is located in the following directory:
$ORACLE_HOME/sysman/config/
The following issues and workarounds apply to the Oracle Enterprise Manager reporting feature.
You cannot access published reports out-of-the-box from the centralized Enterprise Manager Reporting Web Site in release 9.0.2. (Bugs 2196391 and 2205025)
The following manual configuration steps are required after installation in order for the Reporting Web Site to function properly. If these steps are not performed, Enterprise Manager displays the following error when you try to access the Reporting Web Site:
Error URL /em/OEMNavigationServlet was not found on this server
To remedy the problem:
jserv.properties) as follows:
$ORACLE_HOME/Apache/Jserv/etc/
wrapper.env.copy=DISPLAY
wrapper.env=DISPLAY=<full hostname of the reporting web server>
opmnctl stopall
opmnctl startall
If the ServerName property specified in the oem.conf file is different from the Web server hostname that is used to configure reporting, then you will encounter the following error when trying to access the Enterprise Manager Reporting Web Site:
The management server <hostname> is using a different webserver <hostname> for Enterprise Manager reporting services. This webserver <hostname> will be deconfigured and will no longer support Enterprise Manager reporting. To access Enterprise Manager Reports in the future, please use the webserver <hostname>.
To fix this problem, ensure that the ServerName property specified in the oem.conf file located in the $ORACLE_HOME/oem_webstage/ directory is the same as the Web server hostname that is provided to configure reporting with the oemctl configure rws command.
For example, if the ServerName property in oem.conf is foo.oracle.com then when you run oemctl configure rws, be sure to specify foo.oracle.com for the Web server hostname. (Bug 1696527)
In order to configure the Enterprise Manager Reporting Web site, you must run the 'oemctl configure rws' command at the command line on the machine which runs the reporting Web server, and supply the appropriate information as prompted. (Bug 1709809)
If you view Enterprise Manager reports in a non-English environment from Microsoft Internet Explorer release 5.0, you may see garbled, non-legible text in the reports. To workaround this problem, either view the reports from Microsoft Internet Explorer release 5.5 or from Netscape Navigator release 4.7. (Bug 1773959)
When accessing Enterprise Manager reports from a Web browser, you may see a blank page or the following error:
Internal Server Error: The server encountered an internal error or misconfiguration and was unable to complete your request. Please contact the server administrator, you@your.address and inform them of the time the error occurred, and anything you might have done that may have caused the error. More information about this error may be available in the server error log.
To resolve this error:
access_log file, which is located in the following directory:
$ORACLE_HOME/Apache/Apache/logs/
If you find an Xlib error, such as the following, see Section 2.2.5.1:
Xlib: Client is not authorized to connect to Server
java.lang.NoClassDefFoundError: sun/awt/motif/MToolkit
$ORACLE_HOME/Apache/Apache/logs/
If you find an OPM error, such as the following, see Section 2.2.5.2.
[error] OPM: EW: Fails to ping process with pid:11301. Probably the process is hanging and can not response. In most cases, the JVM is either heavily loaded or doing GC. Please increase the jvm timeout value or decrease the jvm heap size to reduce the time of doing GC. The process is about to be restarted.
If an Xlib exception occurred, perform the following steps to fix the problem:
xhost +<host> command on the X Window machine where you are accessing reports from a web browser.
For example, if your reporting web server is running on host servlet1 and the X Window is running on host xserv1, then execute the following command on xserv1:
xhost +servlet1
If your UNIX machine does not meet the hardware requirements for running an X server, see Section 2.2.6.
For example, if your reporting web server is running on host servlet1 and the X Window is running on host xserv1, then execute the following command on servlet1:
If you are using the C Shell:
setenv DISPLAY xserv1:0.0
If you are using the B or K Shell:
DISPLAY=xwerv1:0.0; export DISPLAY
Verify that the X Window environment has been configured properly. For example, execute the xclock command. If you see the xclock window on your X Server (in this case, the reporting Web server), your X Window environment has been configured properly.
apachectl restart
If an OPM error occurred, perform the following steps to fix the problem:
$ORACLE_HOME/Apache/Jserv/etc/
Try setting the value to around 30 or 35.
$ORACLE_HOME/Apache/Jserv/etc/
The following values should be ideal:
wrapper.bin.parameters=-Xms64m wrapper.bin.parameters=-Xmx128m
If your computer does not meet the hardware requirements for running an X server, there is an alternative to the remote X server solution described in Section 2.2.5. The alternative is to run a pseudo-X server such as the X Virtual Frame Buffer, or XVFB.
The main difference between XVFB and the standard X server software is that XVFB uses an in-memory virtual frame buffer instead of a hardware frame buffer. As a result, XVFB can be run on almost any Unix machine, including "headless" middle-tier machines which lack a hardware frame buffer and keyboard. If your UNIX vendor does not provide XVFB binary, then you must build the XVFB binary from the X.Org source code.
The following steps describe how to build XVFB on Solaris 8:
Source code for XVFB is available from the X.Org's FTP server. The X Window System source code is distributed in a number of tar files. Only the first tar file is needed to build XVFB. The necessary tar file for version X11R6.5.1 is available at:
ftp://ftp.x.org/pub/R6.5.1/tars/xorg-1.tar.gz
Download and unzip/untar this tar file on the machine where XVFB is being built. This should create a top-level directory named xc (which stands for X Consortium, the former name of X.org).
site.def.
Configuration information for the build is specified in the site.def file, located in the xc/config/cf/ directory. This file needs to be modified to indicate that XVFB should be built. Also, since it is not necessary to build all of the X Window System, a number of items in the site.def file can be commented out to speed up the build.
Finally, for Solaris, the ProjectRoot should be changed to /usr/openwin, which is the normal location for X-related files on Solaris. Or, simply copy this preconfigured site.def file to xc/config/cf.
The final step is to build the XVFB binary. The following software is required:
/usr/ccs/bin/make should be used
Sun WorkShop 6 is recommended, although gcc can be used.
Once you are ready to build XVFB, go to the xc directory and run the following command:
make World
After the build completes (about an hour), the Xvfb executable should be created in xc/programs/Xserver/Xvfb.
For a sanity check on the XVFB executable, run ldd Xvfb to make sure that the link succeeded. You should see something along the lines of:
libsocket.so.1 => /usr/lib/libsocket.so.1 libnsl.so.1 => /usr/lib/libnsl.so.1 libc.so.1 => /usr/lib/libc.so.1 libdl.so.1 => /usr/lib/libdl.so.1 libmp.so.2 => /usr/lib/libmp.so.2 /usr/platform/SUNW,Ultra-1/lib/libc_psr.so.1
The following command line can be used to launch XVFB:
Xvfb :1 -screen 0 1x1x24
The XVFB binary must be run as root or must have the suid bit set. The :1 argument indicates that the XVFB binary runs as server number 1 (the standard X server typically runs as server number 0). The -screen argument indicates that a single screen of size 1x1 with 24-bit pixel depth should be used. Note, since all images are generated off screen, the 1x1 screen size is sufficient for image generation purposes.
Since XVFB must remain running for the lifetime of the application, perhaps the best approach is to automatically launch XVFB during the machine's boot process. XVFB can then remain running as a background process until the machine is shut down.
Once the XVFB executable is running, client applications can connect to XVFB by setting the DISPLAY environment variable to :1. The :1 setting indicates that server number 1 (the XVFB server) on the local machine should be used for all X graphics operations. When running in Jserv, the DISPLAY environment variable should be set in jserv.properties, for example:
wrapper.env=DISPLAY=:1
If you view Enterprise Manager reports in a Japanese, Simplified Chinese, or other multibyte character set environment, you will see garbled, non-legible characters in the reports. To fix this problem, perform the following steps. The following example assumes the Japanese Solaris environment: (Bug 1701315)
The file is located in the following directory:
$ORACLE_HOME/Apache/Jserv/etc/jserv.properties
wrapper.env=LANG= line and set the appropriate OS locale value.
# Uncomment to set Java default locale based on the LANG environment # variable. #wrapper.env.copy=LANG
For example:
# Uncomment to set Java default locale based on the LANG environment # variable. wrapper.env.copy=ja
By default, the Enterprise Manager Reporting Web Site and browser-based Enterprise Manager use port 3339. However, if port 3339 is already in use by some other component, then during installation and post-installation configuration a different port will be assigned automatically to the Reporting Web Site and to the browser-based Enterprise Manager.
To obtain the new port number, review the portlist.ini file located in the following directory:
$ORACLE_HOME/install/
The portlist.ini file contains a property labeled Enterprise Manager Reporting Port which identifies the port being used by the Reporting Web site.
Discovered Oracle Forms Servers, Oracle Metrics Clients, and Oracle Metrics Servers appear on the Reporting Web Site as ORACLE_FORMS_SERVER, ORACLE_METRICS_CLIENT, and ORACLE_METRICS_SERVER, respectively, and their corresponding product icons are missing.
(Bug 1679949)
Netscape may not display non-English titles or links correctly. If this happens, please set the character set appropriate to your language and select the option Use my default-fonts, overriding document-specified fonts in your Netscape browser. (Bug 1777348)
If you click View Report from the Create Report dialog while using the Enterprise Manager Console on a Sun Solaris computer, the Netscape browser may display the report using a very small text font. To workaround this problem, perform the following steps: (Bug 2265174)
For non-English environments, some charts may display some of the translated messages as square boxes or question marks. (Bugs 2274548, 2199733)
The following release notes apply to the Event and Job Systems available from the Oracle Enterprise Manager Console.
When you create or edit the OC4J Up/Down event, you may receive the following errors:
VTO-2001: Warning! An initialization error occurred. Some data may not be displayed correctly. VTC-1331: The type oc4j is not available. A node with a target of this type must be discovered before editing this job or event.
You can safely ignore these errors. The Up/Down event will still be registered and work successfully. (Bug #2200975)
If your User-Defined event has failed with the following error, the tagged output specific to user-defined events could not be found in standard output:
User Defined Event" failed: No result returned in output:
This may mean your script failed to execute for a variety of reasons. Some things to look into:
As an additional check, log on to the monitored node using the Node credentials used for the event, then make sure the monitoring script runs correctly.
Associating fixit jobs with unsolicited events is only supported by 9i Intelligent Agents.
The following notes apply to registered events:
The event edit was not submitted because no changes were detected for this event.
To workaround this problem, remove notification for the administrator and also perform another edit such as editing the event's description. (Bug 1761277)
Pre-9i Agents do not support the "Day of the Week" or "Day of the Month" options for Event Schedules. When an event contains at least one target running a pre-9i Agent, then only the "On Interval" option for Event Schedules will be supported. However, the Create Event dialog will not disable the selection of the 2 other options -- "Day of the Week" or "Day of the Month". Selection of these 2 other options will result in error "VNI-4022: Bad V1 schedule string" when attempting to register the event. (Bug 1767845)
If you change a scheduled job by only removing notification for an administrator, you will receive the following error: "The job edit was not submitted because no changes were detected for this job" and the change will not be performed. To workaround this problem, remove notification for the administrator and also perform another edit such as editing the job's description. (Bug 1761277)
During event creation, if targets are specified for an event AFTER selecting the fixit job for the event, then event registration against these targets may fail if these are not valid targets for the associated fixit job. If this problem occurs, first modify the fixit job by adding these additional targets. Make sure the job status for the newly added targets is 'Fixit'. Then, in the Registered pane in the Console, select and de-register the event from the targets where registration has failed. When the de-registration against those targets is complete, modify the event by adding back the targets. The event registration against those targets should succeed this time. (Bug 1772485)
If the OS character set when the Intelligent Agent is started up and the database character set which the Agent is managing are different, the "Alert" event returns corrupted characters. This happens if alert.ora file has non-US7ASCII characters. (Bug 1689939)
If you configure a paging carrier with invalid entries and then test that paging carrier, it takes five minutes before a message appears indicating that it failed to send the test page. (Bug 1757910)
Enterprise Manager does not support multibyte characters in both UNIX and NT environment. This is due to the restriction of these platforms. (Bugs 2165255, 2213911)
To register an event for an Oracle9iAS Release 2 (9.0.2) target, the active Intelligent Agent on the target's host must be Version 9.0.2 or higher.
If a schema name contains a "." delimiter, when viewing objects under this schema, their details will appear to be empty. (Bug 1296054)
For pre-9i managed databases, to use the Summary Advisor wizard, you need to have an external procedure agent setup in the listener.ora and tnsnames.ora files. Refer to the Net8 Administrator's Guide for more information.
OS Authentication will work only when connecting directly to a database (client/server mode). OS Authentication is not supported when connecting through the Management Server.
If you get this error while running SQL*Plus Worksheet. The error also indicates that no server side file system access is allowed when SQL*Plus connects to OMS. In this case, a script is attempted to be executed on the OMS machine, where SQL*Plus process is run in restricted mode. For security reasons only client side scripts are allowed through the "Run Local Script..." menu. See the SQL*Plus Worksheet Help for additional information.
In order to use SQL*Plus Worksheet in a language other than English, you must first modify the dbappscfg.properties file located in the DRIVE_LETTER:\ORACLE_HOME\sysman\config\ directory. Modify this file by adding a value for SQLPLUS_NLS_LANG; the value should be the same as the default NLS_LANG parameter.
On a Japanese operating system only: When you insert data into or select from NCHAR/NVARCHAR2 column using "Graphical Select mode" against 8.1.7 database or earlier, you will get an error. You must use FREE SQL mode with 'N' keyword to insert into or select from NCHAR columns.
Example 1: INSERT INTO nchar_tab(ncol1, ncol2)values (N'XXX', N'YYY')
Example 2: SELECT ncol2 from nchar_tab where ncol1 = N'XXX'
When selecting a database from the navigation tree, the Quick Tour button results in the error:
VTA-1002: Quick Tour could not show the URL: /oracle/sysman/qtour/dbapack{0}/dba.htm
A similar error occurs for each of the items under a particular database. The workaround is to launch the quick tours from the main console quick tour menu (Bug 1770792)
When you insert data into or select from NCHAR/NVARCHAR2 column using "Graphical Select mode" against 8.1.7 database or earlier, you will get an error.
You must use FREE SQL mode with 'N' keyword to insert into or select from NCHAR columns.
INSERT INTO nchar_tab(ncol1, ncol2)values (N'XXX', N'YYY')
SELECT ncol2 from nchar_tab where ncol1 = N'XXX'
(Bugs 1546466 and 1549367)
SQL*PlusWorksheet cannot show correct Chinese characters when connected to an OMS (Bug 1686104). The workaround is to specify the following in dbappscfg.properties(Solaris only):
SQLPLUS_LD_LIBRARY_PATH = SIMPLIFIED CHINESE_CHINA.ZHS16GBK, SQLPLUS_LD_ LIBRARY_PATH = $ORACLE_HOME/lib and SQLPLUS_ORACLE_HOME = $ORACLE_HOME
If your Management Server is running on a Solaris node and the Enterprise Manager repository database is running on either an HP 10.20 node or a Siemens SNI node, then you need to set the following SQLNET.ORA parameter on the Management Server node:
DISABLE_OOB=ON
The SQLNET.ORA file is located in the $ORACLE_HOME/network/admin/ directory, where $ORACLE_HOME is the directory where the Management Server is installed. This parameter will prevent sqlnet connection hangs that are caused by differences in the way Sun's and HP's/Siemen's TCP layers handle simultaneous urgent messages on the same socket.
After you set this parameter, you need to stop and re-start the Management Server.
If your repository is in database release 8.1.7.0 or 8.1.7.1, be aware that database base bug 1393049, which causes some SQL queries to return incorrect results, may affect Enterprise Manager in a variety of ways. For example, the bug may manifest itself through Enterprise Manager via incorrect lists of events or jobs in history; incorrect list of event alerts; incorrect list of active jobs, invalid notifications being sent, etc.
To fix the problem, refer to the table below:
To avoid these known database problems entirely, Oracle recommends that you create your repository in a release 9.0.1, 8.1.7.2, or 8.1.7.1 (Sun SPARC Solaris only) database. (Database Base Bug 1393049)
If your repository is in database release 8.1.7.x for any operating system, Oracle recommends that you apply the patch for bug 1733170. Failure to apply this patch may result in intermittent Management Server faults due to segmentation violations. (Database Base Bug 1733170)
To use a release 9.0.2 Management Server with a release 9.0.1 Enterprise Manager Repository, perform the following configuration steps after installation:
When prompted, select Configure local Oracle Management Server; then, choose Edit, and supply the release 9.0.1 Repository connection information.
oemctl import registry $ORACLE_HOME/sysman/admin/OMSIASRegistry.registry
After logging into browser-based Enterprise Manager, the Console will initially be 1/4 its normal size and appear in the upper left corner of the screen. To workaround this problem, resize the Console whenever this occurs. (Bug 1636481)
If you are running browser-based Enterprise Manager from a Netscape browser, do not specify *.<domain> in the manual proxy configuration.
For example, do not specify *.us.oracle.com. Instead, specify .<domain>. For example, us.oracle.com. (Bug 1936133)
Oracle9i Application Server (Oracle9iAS) includes a new version of the Oracle HTTP Server. The following sections describe issues and problems that affect discovering and using previous versions of Oracle HTTP Server with the Oracle Enterprise Manager console.
If the Oracle HTTP Server is configured to listen at a port number below 1024, the you must have root privileges in order to start the Oracle HTTP Listener.
To start and stop the Web server using the Enterprise Manager Console, you need to set the Preferred Credentials for the web server node. If the server is listening at a port number below 1024, the preferred credentials must allow for "root" privilege. Otherwise, you will not be able to start or start the Web server from the Console.
Enterprise Manager parses the HTTP Listener configuration file, httpd(s).conf, linearly. Because of this, directives such as 'Listen', 'IfDefine', 'VirtualHost', 'ResourceConfig', and 'Directory' are not supported (any values specified here will either be ignored or read as variables and will cause the discovery script to fail).
To configure the Web server for discovery by the Enterprise Manager Agent, use the following directives:
'Port' -- used to set the port number of the web server target 'Location' -- used to set the server status URL
Include files are ignored by the Enterprise Manager Web server discovery script. The discovery script doesn't interpret any files in the HTTP Listener configuration file (httpd(s).conf), that contain the include verb.
This section describes known errors in the documentation.
The Enterprise Manager Configuration Guide incorrectly documents that the Enterprise Manager Web Site for UNIX bundles Oracle JInitiator release 1.1.8.16o. Enterprise Manager Web Site for UNIX actually bundles (and consequently browser-based Enterprise Manager uses) Oracle JInitiator release 1.1.8.18o.
Oracle is a registered trademark, and Oracle9i is a trademark or registered trademark of Oracle Corporation. Other names may be trademarks of their respective owners.
Copyright © 2002 Oracle Corporation.
All Rights Reserved.
|
|
 Copyright © 2002 Oracle Corporation. All Rights Reserved. |
|