VMware vCenter Operations Manager Installable 5.8.2 Release Notes
VMware vCenter Operations Manager 5.8.2
Last Document Update: 24 JUNE 2014
Check frequently for additions and updates to these release notes.
These release notes include the following topics:
- What's New
- Before You Begin
- Product Compatibility
- Installation and Upgrade Instructions
- Resolved Issues
- Known Issues
vCenter Operations Manager 5.8.2 is the latest release of VMware's integrated operations suite, converging performance, capacity, and configuration management. This new release introduces the following enhancements.
Enhanced Networking Support
Added VMware vSphere Distributed Switch as a managed object with associated metrics within the NSX Management Pack. vSphere Distributed Switch is compatible with vSphere 5.5.
New ESX Host Property
Exposed the ESX Host UUID as a new property for host managed objects in the VMware adapter.
Enhanced Widget Interaction
The Heatmap Widget is enhanced in the following ways:
- Can be a target of widget interaction
- Contents can be filtered based on the source of interaction.
vCenter Chargeback Manager EOA
As of June 10, 2014, vCenter Chargeback Manager reaches End of Availability (EOA) and is no longer bundled with the vCenter Operations Management Suite.
Apache Struts Update
Updated Apache Struts to remediate several vulnerabilities:
- DoS due to small buffer size (CVE-2014-0050)
- ClassLoader ParametersInterceptor manipulation (CVE-2014-0094)
- ClassLoader ParametersInterceptor manipulation (CVE-2014-0112)
VMware vCenter Operations Manager version 5.8.2 is available as a standalone executable or as a vApp that you deploy with the vSphere Client.
You can install and upgrade the standalone version of vCenter Operations Manager 5.8.2 on the following operating systems:
- Windows Server 2003 64-bit SP2 or later
- Windows Server 2003 R2
- Windows Server 2008 64-bit
- Windows Server 2008 R2
- Red Hat Enterprise Linux 5 64-bit or later
vCenter Operations Manager 5.8.2 supports the following database servers and versions:
- Oracle 10g Release 2
- Oracle 11g Release 2
- Microsoft SQL Server 2005 SP2 or later
- Microsoft SQL Server 2008
- Microsoft SQL Server 2008 R2
For system requirements, installation instructions, and configuration options for the standalone version, see the VMware vCenter Operations Manager Installation and Configuration Guide for Windows and Linux.
For information about vApp deployments, see the VMware vCenter Operations Manager Deployment and Configuration Guide.
Refer to the VMware Compatibility Guide for more information about VMware certified compatibility for vCenter Operations Manager: http://www.vmware.com/resources/guides.html.
The VMware vCenter Operations Manager Installation and Configuration Guide for Windows and Linux lists compatible versions of vSphere and other components that vCenter Operations Manager supports.
vCenter Operations Manager 5.8.2 supports the following browsers:
- Apple Safari 6
- Google Chrome 24 and 25
- Internet Explorer for Windows 8.0 and 9.0
Note: The Compatibility View mode must be disabled.
- Mozilla Firefox 18 and higher
The minimum supported browser resolution is 1024 by 768 pixels.
To install the standalone version of vCenter Operations Manager 5.8.2, run the 5.8.2 installation files for your operating system. For more information, see the VMware vCenter Operations Manager Enterprise Installation and Configuration Guide for Windows and Linux.
This procedure is applicable when updating to vCenter Operations Manager 5.8.2 from the following product versions:
- vCenter Operations Manager Enterprise 5.0, 5.0.1, 5.0.2, 5.0.3, 5.6, 5.7, 5.7.1, and 5.8.x, including any Hotfix versions.
You cannot directly upgrade to vCenter Operations Manager 5.8.2 from the following product versions. You must upgrade from 1.x to 5.0.x before you can upgrade to version 5.8.2.
- vCenter Operations Manager Enterprise 1.0, 1.0.1, 1.0.2, including any Hotfix versions.
- Depending on the operating system, download the latest .exe or .bin file to the machine where the old product version is installed.
- Run the .exe or .bin file.
- Select the Reinstall (Upgrade) option.
- Select the installation mode.
- To preserve your settings from the original installation, do not change the default setting.
- To modify installation options, select Basic or Custom.
- On the Database configurations screen, do not modify any settings, and click Finish.
- On the main installation screen, click Done.
- Check the following log files under the vcenter-ops/user/log folder for upgrade issues:
- If adapters are installed on the system that you upgraded, download and install the latest adapter versions.
- Log in to vCenter Operations Manager (custom interface) and verify that the user interface displays properly.
Note: The database upgrade may fail if you have insufficient privileges. The minimum permissions and privileges required to upgrade the database are:CREATE TABLE, CREATE TRIGGER, CREATE SEQUENCE, CREATE SESSION, and CREATE PROCEDURE.
vCenter Operations Manager 5.8.2 includes fixes for the following issues found in prior releases.
- Datastore disk space allocation
vCenter Operations Manager no longer reports incorrect disk space allocation for VMs sharing the same datastore.
- Default policy access error
Attempts to access the Default Policy under certain conditions, no longer result in an Error 500 message from the server.
- Unlock imported LDAP user
The ability to unlock an imported LDAP user account from the vCenter Operations Manager UI is no longer prevented.
- LDAP group privileges
Adding or removing a user from an AD group no longer fails to provide the user with proper access control after LDAP auto-sync or re-import.
- Side-by-Side upgrade
Side-by-side upgrade from vCenter Operations Manager 5.8 no longer fails.
- Cluster detail values
No longer showing erroneous values or units for the Avg VM CPU and Avg VM Mem properties of a cluster object.
- Policy group deletion
Deleting groups associated with policies that are visible in Folders no longer fails.
This release contains the following known problems.
- Some metrics are missing in the Custom UI
Some metrics listed in the documentation do not appear in the Custom UI. For example, some UI elements that visualize attributes or metrics, hide property-type attributes because of certain limitations related to that type of attribute. These metrics are hidden from the attribute package editor in the Custom UI but may still contribute to the overall attribute count displayed when selecting a group in the attribute package editor.
Workaround: This is expected behavior and there is no workaround.
- Metric names are displayed in English only
Regardless of the language setting, metric names are only displayed in English.
- Problems with attribute packages management with vCenter Configuration Manager
If you select Manage Attribute Packages, and configure VCM as a new instance, you cannot configure vCenter Configuration Manager as a new instance if the entry contains Non-ASCII characters. If you enter Non-ASCII characters, no information displays for the Instance name.
Workaround: Enter ASCII characters.
- After upgrading, the SAP adapter stops collecting metrics
After an upgrade, the SAP adapter stops collecting metrics. During the upgrade, the sapjcorfc.dll file is removed. This file is used to collect data.
Workaround: After the upgrade, the sapjcorfc.dll file needs to be added manually to installation/common/bin.
- 404 error appears during upgrade
When upgrading either the Windows or RHEL version of vCenter Operations Manager from 5.0 or 5.0.1 to 5.8.2 directly or in any chain upgrade path (for example, 5.0 > 5.7.1 > 5.8.1 > 5.8.2), the UI displays a "404 error" and the following NullPointerException error is written to the localhost.XXX.log (tomcat\logs\):
SEVERE: Exception sending context initialized event to listener instance of class com.integrien.alive.ui.util.MainPortalListener java.lang.NullPointerException at com.integrien.alive.ui.util.WebProperties.loadProperties(WebProperties.java:117)
Workaround: Add the following lines to the web.properties file in <installation_directory>\user\conf\web.
#licenseCacheUpdateInterval Defines the license update cache interval
licenseCacheUpdateInterval = 86400000
#textDisplay_root The root folder for the text display widget
#Frequency and time interval for stack trace
timeIntervalForStackTracing = 3600000
#maxMetricsPerAlertToPlot The maximum number of metrics to plot for a consolidated alert in a single pass
#tagPickerMaxResourcesPerNode the maximum number of resources per node displayed in tagpicker
- System performance impacts data collection
A large vCenter Management Operations deployment that manages many vCenter Servers can experience a computational bottleneck when calculating world object group metrics during every data collection cycle.
Workaround: The Analytics virtual machine in vCenter Operations Manager dynamically disables the world metrics calculation. No manual intervention is required.
- In the vCenter Operations Manager custom environment, the datastore count in the Custom Overview displays zero.
When the Datastore tag in the Environment Overview tree view is expanded, all of the listed datastores display datastore count "0". Resources related to the selected datastore are not visible.
- The export to .csv file fails if there is no file name specified in Conf.properties file
If you configure the Conf.properties file to export data to .csv file, and you provide the path of the file, the Repository Adapter does not create the file under the path specified in the conf.properties file.
Workaround: Include a file name with the file path.
A resource that is in maintenance mode in the vCenter cannot be permanently removed from maintenance mode in vCenter Operations Manager.
vCenter Operations Manager pulls maintenance mode data from the vCenter, and puts resources in manual maintenance mode in vCenter Operations Manager. While you can remove a resource from maintenance mode in vCenter Operations Manager, during the next collection cycle, the resource is put back into maintenance mode.
Workaround: There is no workaround. If a resource is in maintenance mode in the vCenter, do not remove the resource from maintenance mode in vCenter Operations Manager.
- NoClassDefFoundError in Tomcat log
Hyperic API Adapter collection does not occur when using the standalone version of vCenter Operations Manager. NoClassDefFoundError appears in the Tomcat log. This is because two JAR files are missing from the directory $ALIVE_BASE/tomcat/webapps/HttpPostAdapter/WEB-INF/lib.
Workaround: Copy the following files from $ALIVE_BASE/common/lib to $ALIVE_BASE/tomcat/webapps/HttpPostAdapter/WEB-INF/lib and restart the web service:
- Unable to uninstall non-English language versions
When attempting to uninstall a non-English language version of vCenter Operations Manager on Windows, all selections on the uninstall menu are grey and inactive.
Workaround: Launch the installer, choose English and re-install using the installation wizard. Then uninstall normally.
- Unable to install vmware-om-adapter on Windows Standalone system
When vCenter Operations Manager is upgraded, the path to the JRE is lost. During the VMware adapter upgrade, an error message appears: "Could not find a valid Java virtual machine to load. You may need to reinstall a supported Java virtual machine."
Workaround: After upgrading vCenter Operations Manager, add '
<VCOPS_BASE>/jre/bin' to the path environment variable and run the installer. For example, add C:\vmware\vcenter-operations\jre\bin to the path environment variable.
- vCenter Server health badge and metric profile change
Changing the metric profile (from Balanced Profile to Full Profile or from Full Profile to Balanced) can affect the active fault score of a vCenter Server object. The fault badge score will be reset to zero (0) from a previous score although the fault alert is still active. This can directly affect the health badge of the vCenter Server object. This is specific only to the vCenter Server objects. Fault badges of other objects are not affected by profile changes.
Workaround: To change Fault and health badge scores to their previous state, restart the collector.
- Some custom interface widgets lose configuration settings after upgrading from 5.0 hotfix to 5.8.2
The Size By, and Color By settings selections highlights for Heat Maps are lost after upgrading, but the Heat Map displays data for the selected options only. The selection highlighting for Metric Weather Map, and Top-N Analysis widgets is lost after upgrading, but the filtering criteria displays the selected metric.
Workaround: Before you upgrade from 5.0, remove the widget configurations for the Heat Maps, Metric Weather Map, and Top-N Analysis widgets.
- The feature to edit the eventlist.txt file settings is not supported in 5.x releases
In vCenter Operations Enterprise 1.0.2, you can edit the eventlist.txt file to specify that certain events can generate alerts. This feature is documented in the vCenter Operations Administration Guide (Custom UI); however, it is not supported in vCenter Operations Manager 5.x.
- The CPU load heat map displays the wrong color
In the custom interface, the CPU load heat map under Heatmaps dashboard > CPU Load widget > Hosts by CPU Usage, displays the color for memory usage instead of CPU usage. This applies to upgrades from vCenter Operations Manager 5.7.1 or 5.7.2 to 5.8.2. It does not apply to a new installation of vCenter Operations Manager 5.8.2.
Workaround: To configure the correct color display for the CPU Load Widget:
- Select the CPU Load Widget and click the edit widget icon.
- From the Edit screen, choose the CPU Usage from the "color by" option and choose Usage (%).
- Click the update selected configuration icon (image of a disk).
- Click OK to save the configuration.
- Custom views and reports are overwritten after upgrade
If you upgrade to vCenter Operations Manager 5.8.2, and your custom views or reports are overwritten, it is because you are using the same view or report key as our out-of-the-box view and report key for 5.8.2.
Workaround: To recover your version of a custom view or report, find the corresponding XML that contains your view or report definition. Change the view or report key and run the "update" command.
- After upgrading, no metrics display in the EMC Storage Dashboard
If you upgrade from vCenter Operations Manager 5.6.1 to 5.7.x , and have EMC VNX adapters, no metrics display in the EMC Storage Dashboard if you are using Firefox.
Workaround: Use Chrome or Internet Explorer instead of Firefox.
- After upgrading to 5.8.2, the Support Dashboard reports the error:
'Not all key information was found....'
When you select Admin > Support > Info > System Tools and click the "Go to Support Dashboard" button, the following error message displays: "Not all key information was found..." This error can occur if you use remote collectors or change the name of a local collector.
Workaround: After the upgrade, edit the widget so that all of the collectors are selected.
- Adapter instance test fails if you click the Test button too soon after installing the adapter
During adapter installation, you click the Describe icon to begin the describe process. The Collector restarts approximately two minutes after the describe process is finished (the "Describe done" status appears in the user interface). If you configure an adapter instance and click the Test button before the Collector has a chance to restart, the test fails.
Workaround: Because the duration of the describe process can vary, wait at least 10 minutes before you click the Test button after you install an adapter.
After upgrading an adapter, the collector.log reports an exception error
After upgrading an adapter on vCenter Operations Manager installations on Widows and Linux operating systems, the collector.log reports the errors:
These errors are caused by old jar files that should have been deleted during the adapter upgrade.
Workaround: Before upgrading the adapter, stop vCenter Operations Manager services, install the adapter, then start the vCenter Operations Manager services. If the duplicate jar files are not deleted, and the collector.log continues to report the errors, repeat the workaround.
Tomcat server does not start after upgrading to vCenter Operations Manager 5.8.2
The Tomcat server fails to start and generates an exception similar to "SEVERE: Exception starting filter struts Unable to load configuration. - [unknown location]" This problem occurs when InstallAnywhere could not remove the old .jar files and there is a conflict between the old and new .jar files.
Workaround: Find the conflicting struts .jar files in $ALIVE_BASE/tomcat/webapps/ROOT/WEB-INF/lib/, remove or rename the old .jar files, and restart the Tomcat server. For example, rename or remove old $ALIVE_BASE/tomcat/webapps/ROOT/WEB-INF/lib/struts2-core-184.108.40.206.jar file to resolve the conflict with a newly installed $ALIVE_BASE/tomcat/webapps/ROOT/WEB-INF/lib/struts2-core-220.127.116.11.jar file.
vCenter Operations Manager custom interface might stop responding after an upgrade
The vCenter Operations Manager custom interface might stop responding about a day after an upgrade. Existing user sessions remain open, but you cannot create new login sessions. Views stop responding and data collection stops. This problem occurs because the upgrade process does not check the database deployment and table space configuration. If a third party made custom table space modifications between updates (for example, to improve performance), the upgrade can place the indexes into the wrong table space and the data table space can exceed its allocated space requirement. This problem can also occur if the default table space becomes full and no modifications are applied between updates.
Workaround: Use one of the following workarounds, depending on the cause of the problem:
- After the upgrade, migrate the indexes to the appropriate table space.
- After the upgrade, revert the database to the default configuration on a larger table space.
Disk Space metrics collection fails after upgrading vCenter Operations Manager
Disk Space metric collection from vCenter Storage Monitoring Service might fail after upgrading from vCenter Operations Manager version 1.0.x to version 5.6. This failure can lead to missing disk space metrics and scores for the Risk and Efficiency badges.
Workaround: Find and remove the file vijava2120100715.jar under $ALIVE_BASE/user/plugins/inbound/vmwarevi_adapter3/lib/ and restart the collector or vCenter Operations Manager services.
vCenter Operations Manager custom interface displays characters of high ASCII as question marks when connected to an Oracle database with the default character setting
For example, if dashboards or resource tags that have high ASCII characters in their names, those letters appear as question marks (?).
Workaround: Configure the Oracle database to use a Unicode character set such as UTF-8 to make characters display correctly.
Exception occurs when you perform a search on the Environment Overview page
The message "Error 'Could not execute query' in the UI, and Exception in the web.log" occurs when you perform a search on the Environment Overview page in the Custom user interface. This problem is caused by a known issue in Oracle 10g.
Workaround: For pre-11gR2 databases, set cursor sharing to 'EXACT' (the default setting). For 11gR2 and later databases, no changes are required and the problem is resolved in 11gR2 Fixpack 2 (18.104.22.168).
Installing vCenter Operations Manager Standalone with MS SQL in console mode on Linux fails to test the database connection
Testing the database configuration will fail due to .RSA and .SF file inconsistencies when the configuration wizard is run from InstallAnywhere, but this failure doesn't impede the overall installation process.
Workaround: This failure can be ignored. The installation is successful afterwards. If the user wants to test the database connection which failed during setup, the configuration wizard can be run from the bash shell using the "$HOME/Configure_VMware_vCenter_Operations -i console" command.
Cyclic redundancy check (CRC) errors manifest as missing data
CRC errors appear as missing data for a given resource.
Workaround: Check the vcenter-ops/data/vcops/log/analytics.log file to see if CRC-related errors are logged, for example:
com.integrien.alive.FSDB.FSDBException: CRC Failure in file for Resource 6773, filename = /data/vcops/data/6/6773/2011_10_6773.dat
2011-11-11 03:04:31,552 ERROR [Calc DT worker thread 3]
com.integrien.analytics.datareceiver.MetricDataRetriever.getDataPointBundle - CRC Failure in file for Resource 6773, filename = /data/vcops/data/6/6773/2011_10_6773.dat
If the log file contains CRC-related errors, set the FSDBCheckAndRepair property to 2 in the vcenter-ops/user/conf/analytics/advanced.properties file. Errors are corrected within seven days.
You must update the statistics of data tables periodically so that the Oracle SQL native query analyzer can operate and fetch data efficiently
The vCenter Operations Manager custom interface does not provide a user interface to configure database updates.
Workaround: See Automated gathering of statistics for Oracle database tables and their indexes (KB 2016451).
DocumentationThe following product changes were not described in the release documentation.
Help error in identifying undersized or oversized machines
For Identifying Undersized Virtual Machines, the VM area taken in account should be documented as the area above the "CPU/Mem Demand greater than" threshold minus the area under the "CPU/Mem Demand greater than" threshold.
For Identifying Oversized Virtual Machines, the VM area taken in account should be documented as the area under the "CPU/Mem Demand less than" threshold minus the area above the "CPU/Mem Demand less than" threshold.
All and group options for dashboard commands
Missing information in the custom interface online help: "Data Age" field displays "Unknown" in the VCOps/VCM Compliance breakdown
vCenter Operations Manager will only display the last collection date (Data Age) if a collection filter is assigned to the VCM compliance rule group.
Missing information on consolidated alerts in the custom interface documentation
See Aggregating and Filtering Alerts in vCenter Operations Manager 5.6 (Custom UI) (KB 2039020) for details.
For the dashboard import, dashboard delete, and dashboard reorder commands, you can use one of the following options instead of specifying a user name:
* --allThe --all option applies the command to all user accounts. The --group:<group_names[,<group_name>] option applies the command to all user accounts that belong to the specified user groups. For example, the command dashboard import group:Administrators,Operators "templates\MyDashboard.bin" imports the dashboard MyDashboard.bin to all user accounts that belong to the Administrators and Operators user groups. The dashboard import command also has a --set <rank> option that enables you to set the dashboard order. For example, the command dashboard import MyUser2 "dashboards/MyDashboard.bin" --set 1 imports the dashboard MyDashboard.bin and makes it the first dashboard in the dashboard list for the user MyUser2.