Updated on 25 SEP 2018
VMware vRealize Operations for Horizon 6.5 | 21 SEP 2017
Check for additions and updates to these release notes.
What's in the Release NotesThe release notes cover the following topics:
- About VMware vRealize Operations for Horizon
- What's New
- Upgrading to Version 6.5
- Resolved Issues
- Known Issues
VMware vRealize Operations for Horizon extends the functionality of VMware vRealize Operations Manager to monitor and manage VMware Horizon environments. vRealize Operations for Horizon collects data about your resources and presents that data in preconfigured dashboards for both real-time and predictive analysis.
The vRealize Operations for Horizon dashboards have been expanded for the following use cases:
- The Horizon Overview dashboard provides the administrator an overall view of their Horizon and NVIDIA deployment. In addition to performance metrics for the Horizon pod, users can also gain insight into NVIDIA GRID trends as well as alerts from NVIDIA GRID.
- The Horizon Help Desk dashboard provides session-specific GPU utilization, GPU memory usage and encode metrics, GPU type, and user-specific alerts from GRID.
- The Horizon End User Experience dashboard includes a “weather map” or “heat map” for GPU use cycle and memory, warning administrators of performance degradation before users are impacted.
- NVIDIA GRID vGPU Monitoring: You can now monitor the usage of NVIDIA vGPUs on ESXi hosts, including vGPU usage per session as well as the usage and health of each vGPU in your environment. To use this feature, import the vGPU dashboards after installing vRealize Operations for Horizon.
- Additional Aggregation Metrics on User and Pool Level: vRealize Operations for Horizon provides out-of-the-box metrics on usage information at the user and pool level, including historical records of desktop pool access for a specified user. You can download the report in CSV format.
- Cloud Pod Architecture Support: vRealize Operations for Horizon provides more information on the cloud pod architecture. You can see whether a pool is a global pool, whether a global entitlement is assigned to the pool, and whether a session is a global session.
- New Design for Logon Time: vRealize Operations for Horizon provides a new way to calculate logon time. This feature is only available with newer versions of Horizon.
- New Process Alerts: vRealize Operations for Horizon helps you monitor processes by triggering an alert and offering a recommendation for any process abnormalities in your environment.
vRealize Operations Manager for Horizon 6.5 requires the following VMware products:
- VMware Horizon (with View) 6.2.4 to 6.2.7 or VMware Horizon 7.0 to 7.4
- VMware vRealize Operations Manager 6.4 to 6.6.1
vRealize Operations Manager for Horizon is also compatible with the following products:
- VMware App Volumes 2.12.x, 2.13.2, and 2.13.3
- VMware Unified Access Gateway 2.9 and 3.1
For the most current information about supported versions, see the VMware Product Interoperability Matrices.
This section contains notes about the upgrade process. For complete upgrade instructions, see Upgrading vRealize Operations for Horizon.
- You can upgrade from vRealize Operations for Horizon 6.1 or later to vRealize Operations for Horizon 6.5. If you have vRealize Operations for Horizon 6.0 or earlier, you must first upgrade to vRealize Operations for Horizon 6.1 and then to vRealize Operations for Horizon 6.5.
- Dashboards, views, and reports from older versions are retained during the upgrade process. These legacy items are prefixed with the word View and can be removed safely.
- Due to an existing limitation, you must restart the vRealize Operations Manager cluster after the upgrade for the process to complete.
- Sessions that are active during the upgrade might not report data after the upgrade is complete. Restart the desktop agent for any active sessions that are not reporting.
- After you upgrade the broker agent, you must pair it with the Horizon Adapter instance again. All other broker agent configuration data is retained.
The vRealize Operations for Horizon user interface and documentation are available in English, Japanese, French, German, Simplified Chinese, Traditional Chinese, Korean, and Spanish.
vRealize Operations for Horizon has the following limitations.
- Physical desktops in a pool are not monitored.
Physical desktops included in a desktop pool are not monitored or shown in the user interface. However, physical RDS hosts are monitored and shown in the user interface.
- High availability (HA) failover can result in the Horizon Adapter not collecting data.
For more information, see KB 2092874: Best practices for configuring vRealize Operations for Horizon when HA is enabled.
- Using separate network cards for different subnetworks in vRealize Operations Manager is not supported. This can cause the Horizon Adapter instance to bind to an incorrect IP address.
- The adapter configuration files
/usr/lib/vmware-vcops/user/plugins/inbound/V4V_adapter3/work/msgserver.propertiesmust be updated at the same time.
- Existing adapter instance configuration files are not removed during the upgrade process. If the TLS configuration has been modified on the adapter side, the broker and desktop agent might fail to connect. Ensure that the adapter instance and agents use the same TLS configuration.
- For sizing and scaling guidelines, see KB 2098051: vRealize Operations for Horizon sizing guidelines.
- Computer-based AppStacks are not supported. Metrics can be displayed for user-assigned AppStacks only.
- Some users are not listed in reports.
- Session is displayed slowly on dashboards in an instant clone pool.
- Missing session in Cloud Pod Architecture in Horizon 7.2.
- Alert not triggered when the connection server service is down.
- Some relationships are not removed after session logoff.
- The pool entitlement query can be disabled in broker agent.
- Adapter installation is improved so it can be installed faster.
- The Blast protocol metric is not available on the Horizon RDS Pools dashboard.
Blast protocol data for RDS pools is not displayed.
Workaround: Create a super metric to calculate Blast data with the pool object.
- Duplicate App Volumes Manager installations can be added in the broker agent.
In the vRealize Operations for Horizon broker agent configuration utility, you can add App Volumes Manager installations by IP address or by FQDN. It is possible to add an IP address and an FQDN that refer to the same App Volumes Manager installation.
Workaround: In the broker agent configuration utility, delete one of the references manually and restart the broker agent service.
- A large number of application crash alerts might be displayed on the Alerts page.
Desktop agents send information about application crashes to the Horizon Adapter, which generates a notification alert that is displayed on the Alerts page. These alerts are not cleared automatically and can cause the Alerts page to be overpopulated.
Workaround: Set a daily
cronjob to run the
cleanallstopobjects.jarscript. The syntax for this script is as follows:
java -jar cleanallstopobjects.jar vrealize-ip username password records expiration
- vrealize-ip is the IP address of the vRealize Operations Manager instance
- username is a vRealize Operations Manager administrator account
- password is the password for that account
- records is the number of records per page
- expiration is the number of days after which records will be deleted
- Deleted objects might be displayed in the Unified Access Gateway widget.
The Unified Access Gateway widget might continue to display objects that you deleted by using the broker agent.
Workaround: Stop the adapter instance and remove the affected objects from the inventory tree. Then start the adapter instance again.
- Pod names are not displayed in the Horizon Connected Sessions widget within the expected time of 15 minutes.
On the Horizon Help Desk dashboard, the Pod Name column in the Horizon Connected Sessions widget is not populated within 15 minutes even though data is displayed in all other columns.
Workaround: None. Pod names should be displayed within 60 minutes after session launch.
- The desktop agent does not operate properly on a machine with an account that has an AppStack assignment.
When you install the desktop agent on an account with an AppStack assignment in App Volumes, the agent might be installed on the mounted AppStack instead of local storage. In this case, after you restart the virtual machine or log in as another user, the agent will no longer operate properly.
Workaround: Do not install the desktop agent on an account that has an AppStack assignment. You can install it on the base template of the desktop machine or on another account.
- Horizon pods continue to appear on Horizon dashboards even after they are no longer monitored.
You choose to stop monitoring a Horizon pod, but information about that pod still appears on dashboards.
Workaround: Restart the service by running
service vmware-vcops --full-restart. Alternatively, in the vRealize Operations Manager user interface, manually delete the target Horizon pod resource and all Horizon Connection Server, security server, and pool resources related to it.
- Incorrect data is displayed for some entries in the Horizon Connected Sessions widget on the Horizon Help Desk dashboard.
Incorrect data might be displayed for reconnected sessions.
Workaround: Click the Edit icon on the widget and select All Active Sessions.
- After vRealize Operations for Horizon is upgraded, logon-related metrics for existing sessions are not retained.
The following metrics are affected:
- Logon time
- Average logon time
- Maximum logon time
- After the name of a Horizon Adapter instance is changed, related data cannot be displayed on the Horizon Adapter Self Health dashboard.
The Horizon Adapter collector resource does not automatically update to reflect changes to the name of the Horizon Adapter instance.
Workaround: Do not change the name of a Horizon Adapter instance after its creation. Revert any affected instances to their original names.
- Certain network alerts are not generated for user sessions that were established before the Horizon Adapter was configured.
The following alerts are not generated:
- PCoIP|Round Trip Latency (ms)
- PCoIP|Packet Loss Percent
Workaround: None. Alerts will be generated normally after the sessions are disconnected and re-established.
- User names that include multibyte characters might not be displayed correctly.
In the vRealize Operations Manager user interface, user names that include multibyte characters (such as Chinese) might not be displayed correctly.
Workaround: None. Avoid using such characters if possible.