VMware

VMware vCenter Operations Enterprise 1.0.2 Release Notes

VMware vCenter Operations Enterprise 1.0.2 | Build 600724

Last Document Update: 15 February 2012

Check frequently for additions and updates to these release notes.

These release notes include the following topics:

What's New

The following information provides highlights of some of the enhancements in this release.

New System Alert for vCenter Server Disconnects

vCenter Operations Enterprise generates a new system alert "Host <name> is disconnected" when the host disconnects from vCenter Server. vCenter Operations Enterprise cancels the alert when the host reconnects to vCenter Server.

Configurable Alerts for vCenter Server Events

You can specify whether certain vCenter Server events can generate alerts by editing the <vcenter-ops>\user\plugins\inbound\vmwarevi_adapter3\conf\eventlist.txt file. The eventlist.txt file defines the list of events that the vCenter Adapter collects from vCenter Server. To specify that an event can generate an alert, mark the event as an alert by using the following format:

<eventId> = alert

For example:

vim.event.AlarmActionTriggeredEvent = alert

When the vCenter Adapter collects an event that is marked as an alert, it sends a notification message to the vCenter Operations Enterprise server and the Analytics service generates an alert.

Definable Metric Sets for the Metric Graph and Metric Sparklines Widgets

You can use an XML file to specify metric sets for the Metric Graph and Metric Sparklines widgets. A metric set defines a fixed set of metrics for the widget to draw when you select a particular resource. You can select metric set definition XML files from the new Res. Interaction Mode menu when you edit the widget configuration. For step-by-step instructions, see VMware KB article 2011714 (http://kb.vmware.com/kb/2011714).

Upgrade Instructions

To upgrade from vCenter Operations Enterprise 1.0 or 1.0.1 to 1.0.2, run the 1.0.2 installation files. For system requirements and additional information, see the VMware vCenter Operations Enterprise Installation and Administration Guide.

Prior Releases

To view release notes for the prior releases of vCenter Operations Enterprise, see VMware vCenter Operations Enterprise 1.0.1 and VMware vCenter Operations Enterprise 1.0. Known problems from prior releases of VMware vCenter Operations Enterprise are described in the release notes for each release.

Resolved Issues

The following issues are resolved in this release.

  • For virtual machines that have less than 256MB of configured memory, vCenter Operations Enterprise miscalculated the virtual machine memory workload value. Now, vCenter Operations Enterprise compensates for these virtual machines and reports the correct memory workload value.
  • When you configured colors for the Heat Map widget, you could not specify negative values for the minimum and maximum color values. Now, you can type negative values in the Min. value and Max. value text boxes on the Edit widget window when you configure the Heat Map widget.
  • The MIB file for the SNMP Trap Sender (<vcenter-ops>\user\plugins\outbound\snmp_alertplugin\mibsIntegrien-TRAP-MIB.mib) did not include entries for Classic Alerts. Now, Classic Alert entries have been added for KPI HT Breach, Notification, and Abnormality conditions. The fields are the same for each alert subtype.

Known Issues

This known issues for this release are grouped by type in the following sections.

Installation and Upgrade Issues

  • The Password field is not visible on the Alert Handler window when you configure Email outbound alerts
    If you select Email from the Outbound Alert Type drop-down menu on the Alert Handler window, the Password field is not visible because the window is too small. This problem occurs in a fresh install of vCenter Operations Enterprise 1.0.2.
    Workaround: Use the Tab key and Shift+Tab key sequence to move back and forth through the visible and non-visible fields.
  • Fields are missing on the Alert Handler window when you configure email outbound alerts in an upgraded vCenter Operations Enterprise 1.0.2 instance
    If you upgrade from vCenter Operations Enterprise 1.0 or 1.0.1 to vCenter Operations Enterprise 1.0.2, the IS_SECURE_CONNECTION, SECURE_CONNECTION_TYPE, IS_REQUIRES_AUTHENTICATION, USERNAME, and PASSWORD fields are missing on the Alert Handler window when you select Email from the Outbound Alert Type drop-down menu.
    Workaround: Run a script to add the missing fields to the database. See VMware KB article 2012638 (http:/kb.vmware.com/kb/2012638) to download scripts for Oracle and SQL databases. Alert handler instances that were created before running the script are not updated and must be deleted and recreated. After running the script, you might notice that one or more fields are not visible in the Alert Handler window. You can use the Tab key and Shift+Tab key sequence to move back and forth through the visible and non-visible fields.
  • The JVM PATH is missing after the vCenter Operations Enterprise installation and the adapter installer fails
    During the vCenter Operations Enterprise installation, the PATH environment variable is appended with the location of JRE shipped with vCenter Operations Enterprise, adapter installer fails if JRE information is not set in PATH variable.
    Workaround: If the Path variable is not set, execute the following to install the adapter:
    <Adapter Installer>.exe LAX_VM "%ALIVE_BASE%\jre\bin\java.exe"
  • You can install vCenter Operations Enterprise only with root user credentials
    Workaround: Contact VMware Support or Professional Services for information about installing vCenter Operations Enterprise on Linux using non-root user account .
  • If you select HTTPS during installation, the host IP address defaults to localhost
    During vCenter Operations Enterprise installation, if you choose "HTTPS" as the protocol for communication with the Web site, the detected Host IP is replaced with localhost.
    Workaround: Manually replace localhost with the detected IP address before continuing.

Configuration Issues

  • vCenter Operations Enterprise can be set up to be accessed via HTTPS protocol only on the standard port 443.
    Workaround: None. If you want to use HTTPS, you must use port 443.
  • Manually created resources should have a unique identifier
    Manually created resources having the same unique identification (Adapter Kind + Resource Kind + Resource Name or Adapter Kind + Resource Kind + List of Identifiers) can result in wrong system behavior, including resources hanging in start and stop status. This issue also applies to Application Tiers sharing the same name. This does not apply to resources added in system through manual or automatic discovery as uniqueness is enforced by the system.
    Workaround: If resources with the same unique identification are duplicate, delete the duplicate resources. If they are different resources, rename the resource or change the List of identifiers conflict. For Application Tiers, use the following rule for naming: Tier Name=Application Name + ' ' + Tier Type. For example, for "Online Banking" Application tiers, use the names "Online Banking DB," "Online Banking App," and so on.
  • The replication server does not restart as expected
    If ActiveMQ on the replication server side is stopped and then restarted, the replication process does not restart as expected.
    Workaround: When the connection to the replication server resumes, cancel the alert and then restart the analytics when ActiveMQ is available.
  • vCenter Operations Enterprise does not support Network Area Storage (NAS) for storing FSDB files.
    Workaround: Use only SAN or DAS devices to store FSDB files.

User Interface Issues

  • The bread crumb path is lost when clicking objects from the VM details page
    If you click on object on the VM details page, the bread crumb path is lost and shows only the current page.
    Workaround: None.
  • The Adapter Instance Resources status icon displays after deletion
    If multiple Adapter Instance Resources feed data to a resource and an adapter instance is deleted, the Environment Overview page displays the status icon for the deleted adapter instance resource.
    Workaround: None
  • On the Alert Overview page, you cannot search the Root Cause or Info columns
    If the text you enter in the Search field on the Alerts Overview page exists only in the Root Cause or Info column, the search will not find it.
    Workaround: None
  • You cannot remove the Time Range constraint in the Manage Alert Watch List window by clicking the check box next to it
    Workaround: To remove the constraint, bring up the Pick a Time Range pop-up window and remove the date and time from the text boxes.
  • Role-based security is implemented on the server side, which means access rights are checked when an action is taken
    As a result, some actions will not give a message of insufficient access rights until after the user has completed entering information for the action. For example, when cloning a dashboard, the insufficient rights message will not appear until after the user enters the name of the new dashboard.
    Workaround: None.
  • A blank page appears when you log in to vCenter Operations Enterprise if the Skype plug-in for Internet Explorer is enabled
    If you have the Skype Plug-in for Internet Explorer installed and enabled, the browser displays a blank white page after you sign into vCenter Operations Enterprise.
    Workarounds:
    • Disable or remove the Skype plug-in.
    • Press F5 after signing in to reload the vCenter Operations Enterprise page.
  • The Configuration Overview widget reflects only collecting resources
    All of the information shown on the Configuration Overview widget--resource counts, metric counts, and so on--includes only resources for which metrics are being collected. If collection is off for a metric, its data is not included. For performance reasons, the information is taken from the memory used by the analytics process. If metrics are not being collected for the resource, the analytics process does not keep its information in memory.
    Workaround: Start collection for all resources you want to be included in the information shown in the Configuration Overview widget.
  • Maintenance schedules cannot span for more than 24 hours
    Workaround: Split any maintenance schedule that spans two days into two different maintenance schedules.
  • Metric value time stamps are not always displayed
    The time stamps for metric values are not shown when you are viewing a period longer than 24 hours even if Show Data Values is selected for the graph.
    Workaround: To show the time stamp for displayed metric values, view a period of less than 24 hours.
  • vCenter Operations Enterprise stops responding if you delete an application with a large number of resources
    Deleting an application which comprises a large number of resources might take a very long time, and vCenter Operations Enterprise might stop responding.
    Workaround: Delete the individual tiers in the application before you delete the application.
  • Manual fingerprint preview mode time ranges may overlap
    Depending on the screen resolution and number of intervals chosen, creating a manual fingerprint may result in overlapping time range in preview modes.
  • Filtering the Alerts widget by Time Range displays an incorrect message
    When the Alerts widget is configured to filter by Time Range, clicking the Display Filtering Criteria icon in the toolbar of the widget incorrectly displays a message saying There is no selected filter criteria.
    Workaround: To see the active filtering criteria, edit the widget. Each active filter type will have a check mark by it.
  • Filtering the Resources widget by Adapter kind displays an incorrect message
    When the Resources widget is configured to filter by adapter kind, clicking the Display Filtering Criteria icon in the toolbar of the widget incorrectly displays a message saying There is no selected filter criteria.
    Workaround: To see the active filtering criteria, including the selected adapter kinds, edit the widget.
  • The Tag Selector widget state is not saved
    If you leave a dashboard which includes the Tag Selector widget, then return to it, any tags selected for interaction with other widgets will not be marked.
    Workaround: None
  • Environment Overview page and Resources widget do not always sort correctly by health
    When sorting by health, resources may be listed in the wrong order on the Environment Overview page and the Resources widget. This is for two reasons: 1) The resources are sorted according to the value in the database, while the value shown comes from the analytics cache. This means the sorting will be wrong while analytics is loading resources. 2) The database stores only the health state (Normal, Degraded, and so on), not the numeric health score. This means that resources in the same state are not necessarily in numeric order.
    Workaround: None
  • Filtering Environment Overview resource list by health state can list wrong resources
    When filtering the Environment Overview page by health (for example, listing only those resources with "Good" health), the resources listed do not always show as having the requested state. This is because the filtering is done according to the health state in the database, while the health state shown is taken from the analytics cache. if these disagree, the filter appears to include the wrong resources.
    Workaround: None
  • Cannot see total number of metrics for a resource that is not collecting
    If metric collection is stopped for a resource, the total number of metrics for the resource shows as "?." This is true everywhere that the total number of metrics is displayed. This is because, for performance reasons, the number of metrics is taken from the memory used by the analytics process. If metrics are not being collected for the resource, the analytics process does not keep its information in memory.
    Workaround: To see the total number of metrics for a resource, start collection for the resource.

Login and Permission Issues

  • The number of allowed login attempts is incorrect
    vCenter Operations Enterprise allows one more login attempt than is specified in the Allowed Login Attempts field of the Account Lockout Policy pane.
    Workaround: Set the Login Attempts field to one less than the actual desired limit.
  • The maximum length for a credential password is 30 characters
    When entering a credential, you can enter only up to 30 characters for the password.
    Workaround: Make sure that all user names used in credentials have passwords of 30 characters or less.
  • Users without permission to edit adapters or applications can edit them through Environment Overview
    Users with permission to edit resources can edit all resources through the Environment Overview page, even if it conflicts with other security settings. Users without permission to edit adapter instances, but with permission to edit resources, cannot edit adapter instances through the Manage Adapter Instances window, but can through Environment Overview. The same is true for applications. Users who have permission to edit resources, but not applications, are able to edit applications through Environment Overview.
    Workaround: Give permission to edit resources only to users who should be able to edit all resources, including adapter instances and applications.
  • vCenter Operations Enterprise does not observe folder permissions set in vCenter Server
    When you block inheritance of permissions to a folder under the VMs and Templates view in vCenter Server, vCenter Operations Enterprise can still discover and collect from the objects in that folder. For example, if you set the permission on a folder that contains resources to no access and then delete the resources in vCenter Operations Enterprise, the resources later reappear in the VC Relationships widget.
    Workaround: Apply permissions to both the VM and Templates and Hosts and Clusters views to ensure that permission inheritance is blocked.

Linux Issues

  • On Red Hat Enterprise Linux release 5, the services do not restart automatically after you make changes in the configuration wizard
    If you run the configuration wizard in console mode under Red Hat Enterprise Linux release 5 to change properties on an existing server, your changes do not take effect automatically after they are saved.
    Workaround: You must restart the vCenter Operations Enterprise services manually. Use the run level script vcops.sh to restart the services.
  • You cannot uninstall the vCenter Operations Enterprise Collector using console mode on 32-bit Red Hat Enterprise Linux
    The command to uninstall the vCenter Operations Enterprise Collector does not work from console mode on 32-bit Red Hat Enterprise Linux.
    Workaround: None.
  • "Too many files open" message on Linux server
    At times, the vCenter Operations Enterprise Web client can cause Tomcat to have too many file descriptors open at once. This leads to a "Too many files" open error message.
    Workaround: The number of descriptors is limited at both the system level and shell level. You may need to increase the shell ulimit maximum or the maximum number of system file descriptors by adding or modifying the fs.file-max parameter in the /etc/sysctl.conf file.

Documentation Issues

  • The new system alert "Failed to Connect to Replication MQ" is not documented
    This alert occurs when vCenter Operations Enterprise cannot connect to the replication MQ. To resolve this alert, perform the following procedure:

    1. View the replication.properties file and verify that the replication MQ is properly configured.
    2. Verify that the ActiveMQ service configured in the replication.properties file is running. Use JConsole to connect to port 1099.
    3. Verify that the connection to the MQ host on port 61616 and 1099 is available from the machine that is running the vCenter Operations Analytics service.

  • The description of the Root Cause Ranking widget is not accurate
    The Anomalies Activity Period field has been removed from the widget. In Symptom mode, the widget shows root cause information based on anomalies that occurred during the previous six hours, unless a different time period is passed to it by a providing widget. In Resource mode, it shows information based on the previous two hours. When a time period is passed to it by a providing widget, it shows information for all anomalies active in that time period.
  • The description of the Root Cause Ranking pane of the Resource Detail page is not accurate
    The Anomalies Activity Period field has been removed from the Root Cause Ranking pane. In Symptom mode, the pane shows root cause information based on anomalies that occurred during the previous six hours. In Resource mode, it shows information based on the previous two hours.
  • New features in vCenter Operations Enterprise 1.0.2 are not documented
    See What's New for information about the new features in vCenter Operations Enterprise 1.0.2.