VMware

VMware vCenterTM AppSpeed 1.2 Release Notes

VMware vCenter AppSpeed 1.2 | 15 Dec 2009 | Build 41671

Last Document Update: 5 Jan 2010

These release notes cover the following topics:

VMware vCenter AppSpeed 1.2 Features

VMware vCenter AppSpeed provides proactive performance management and service-level reporting for applications running within virtual appliances. By analyzing the traffic that flows between end-users, Web applications, and back-end servers, AppSpeed provides visibility into multitier applications and enables rapid identification of performance issues originating from inadequate resource allocation and application problems.

The new AppSpeed features include:

  • Enhanced AppSpeed scalability AppSpeed scalability has been enhanced, enabling a single AppSpeed Server to monitor significantly larger environments (more servers, applications and transactions). You can now specifically exclude servers that do not need to be monitored.
  • Central management for multiple AppSpeed Servers You can now install more than one AppSpeed Server and control the environment of each AppSpeed Server from a single user interface. This enables additional scalability for monitoring large environments, while delivering centralized visibility of application performance and service levels.
  • Enhanced mapping capabilities Improvements have been made in the mapping algorithms, enabling smarter, more accurate application grouping and transaction analysis.
  • Topology editing from the user interface You can now merge, split or delete applications, providing greater control over how your monitoring data is presented. All of the operations are accessible from the AppSpeed user interface.
  • Enhanced security using a vCenter-based authentication model Connection to AppSpeed Servers is now authenticated against vCenter. A new AppSpeed privilege has been introduced and AppSpeed access is now only granted to AppSpeed-privileged users.
  • New latency-focused Analysis views Two new latency-focused analysis views have been introduced in the AppSpeed user interface, enabling more detailed visibility for latency analysis and troubleshooting:
    • A view that compares a single application or transaction latency between servers
    • A view that compares transaction or server latency with their latency baselines, pointing to latency anomalies

Other VMware vCenter AppSpeed 1.2 Documentation

VMware vCenter AppSpeed Installation and Upgrade Guide

VMware vCenter AppSpeed User's Guide

System Requirements for VMware vCenter AppSpeed 1.2

AppSpeed is supported for the following ESX and vCenter versions:

  • ESX, version 3.0x, 3.5x or 4.x
  • vCenter Server, version 2.5x or vSphere 4.0

In addition, the successful installation of the AppSpeed virtual appliance and plug-in, and the deployment of the probes, requires that the following prerequisites be met:

  • Administrator privileges on vCenter Server, for the installation process
  • SSL private keys for HTTPS sites that AppSpeed is to monitor
  • Administrator privileges on the vSphere Client machine, to install Adobe Flash Player 10

The following system resource requirements must be met:

  • Server
  • 4GB RAM
  • 2 vCPUs

  • Note: Virtual CPUs are created as part of the AppSpeed deployment. Two vCPUs require no more than the equivalent resources of two physical CPUs.
  • Probes
  • 5GB HDD
  • 0.5GB RAM
  • 2 vCPUs

  • Note: Virtual CPUs are created as part of the AppSpeed probe deployment. Two vCPUs require no more than the equivalent resources of a single physical CPU because AppSpeed configures resource limits to ensure that no more resources than are available on one physical CPU are required at any time.

Installation Notes

Known Issues with VMware vCenter AppSpeed 1.2

Server Installation

  • You can only monitor the vCenter environment on which AppSpeed Server is installed. Similarly, with linked vCenter environments, each environment must have an AppSpeed Server installed so that it can be monitored.
    AppSpeed Server must reside on the vCenter on which monitoring is to be performed.
  • You cannot specify the vCenter name in setup. You must specify the vCenter IP address.
  • Harmless FAILED messages are displayed when starting up an AppSpeed machine.
  • When you run setup for the first time, the RTNETLNIK: answers: network is unreachable message might appear.
    The message has no effect on the setup.
  • In rare cases, AppSpeed Server starts up without a network connection meaning that the "Connect" check box is not selected in the virtual appliance settings, and there is a blank network label.
    Manually select the network label and the Connect check box.
  • When reinstalling AppSpeed Server, there is a long delay between the time that you first press ENTER and the time that the first question appears.

UI Installation

  • You must have Administrator privileges on your Windows machine to install the AppSpeed plug-in the first time.

Probe Deployment

  • Sufficient disk space is not automatically verified before deploying probes.
    When there is insufficient disk space for the probes to be deployed, an internal error message is displayed and the Recent Tasks pane displays Insufficient disk space.
  • When you attempt to deploy a probe on a disconnected datastore, the Invalid configuration for device '1' message appears.
    You can only deploy probes on a connected datastore.
  • When you attempt to deploy a probe on a disconnected host, an Internal error message appears (without additional details).
    You can only deploy probes on a connected host.
  • In some cases, two folders are created for a deployed probe, instead of one.
    Do not delete either of the folders.
  • Probes can be started manually from the vSphere Client inventory, before their deployment is complete. This causes the probes to deploy incorrectly.
    Probes start automatically once deployed. You do not need to start them manually.
  • When a probe is deployed, if the specified management network name contains non-ASCII characters, probe deployment fails.
  • An AppSpeed probe template snapshot is automatically created for probe deployment.
    You can remove the snapshot at any time. VMware recommends that you delete the snapshot after the initial probe deployment is complete.
  • After a probe is deleted, its directory sometimes remains in the datastore.
    The deleted probe's directory and contents can be removed.
  • In vSphere, if you use the Enable individual virtual machine automation levels feature, the AppSpeed probe is not attached to the host, and is subject to VMware Distributed Resource Scheduler and VMware High Availability. The probe can move if automatic migration using VMotion occurs, in which case the probe will not be able to monitor the correct traffic.
  • In vSphere, if you disable the Enable individual virtual machine automation levels feature for a cluster, the probes' coverage of traffic can be unpredictable.
  • In some cases, the Failure due to malformed requests from the server message appears when probe deployment fails.
    Delete the failed probe and redeploy it. If the probe deployment freezes, restart AppSpeed from the server console and try again.
  • If probe deployment fails, you might have to manually remove the temporary linked-clone virtual machine that was created during the deployment process.
  • Probe deployment might fail during multi-probe deployment, with the following error: Operation is not allowed in the current state.
    Delete the probe and deploy a new one.
  • If a probe deployment fails, the failed probe is not automatically deleted.
    Select the probe in the Probes Management window of the AppSpeed Administration module and click Delete Probe.
  • Distributed Power Management (DPM) might not work on hosts that contain AppSpeed probes.

User Interface

  • Supported screen resolutions for the AppSpeed user interface are: 1024 by 768 and higher.
    Lower resolutions are only partially supported.
  • The AppSpeed icon appears when you log in to ESX using the vSphere Client although AppSpeed is only available when you log into vCenter. It will also appear if you use vSphere Client to log in to other vCenter Servers that do not have AppSpeed installed.
  • Connection via proxy with authentication is not supported.
  • In some cases, the Probe Management view is not loaded.
    Restart the vSphere Client.
  • Column names are truncated in the Real Time States graph.
  • The list of virtual machines and applications displayed by AppSpeed reflects the current mapping status, regardless of the selected time frame.
  • The following message might appear in the user interface: Attempted to read or write protected memory.
    The message relates to high memory consumption. Reopen the vSphere Client.
  • In some cases, you cannot save export data in CSV format.
    Ensure that the file name has a .csv extension specified.
  • In some cases, limited functionality might occur related to time frame selection. For example, time prior to AppSpeed Server installation cannot be selected.
    Use the Last/By option to specify a time frame as close as possible to the time frame you want to display.
  • If you connect to the AppSpeed user interface for the first time, using a port other than port 80, an error will be returned.
    In the vSphere Client, go to Administration > vCenter Server Settings > Ports and change the HTTP port value to 80.
  • Manually editing the topology using Topology Management might take some time.

Mapping

  • SSL connections that use the Diffie-Hellman key-exchange algorithm cannot be monitored. In the case that only the Diffie Hellman key-exchange algorithm is used by an application or transaction, that application or transaction cannot be mapped.
  • EXCHANGE applications are partially monitored.
    Only Outlook direct connections to an EXCHANGE server are monitored.
  • AppSpeed might not always correctly map and monitor virtual machines that have fault tolerance enabled.
  • During the mapping process, object names might change several times.

Interoperability

  • IP address changes might compromise AppSpeed Server operations (changes due to DHCP, or any other cause).
    If the AppSpeed Server IP address changes, manually delete all probes from the vCenter inventory, run setup in the AppSpeed Server console and redeploy the probes. All data and history are lost.
  • AppSpeed Server time is synchronized with the ESX host.
    VMware recommends that you synchronize the AppSpeed Server ESX host using NTP, to maintain consistent time. If AppSpeed Server is installed in a cluster, and is subject to VMware Distributed Resource Scheduler, ensure that all hosts in the cluster are synchronized using NTP.
  • If AppSpeed Server time is not synchronized, gaps might be encountered in the data monitored by the AppSpeed probes.
  • When you use VMotion to migrate an AppSpeed Server to hosts with different time settings, time jumps might occur in the AppSpeed Server.
    The time jumps might result in inaccurate monitoring data for that time. If time moves backward, some data are erased. If time moves forward, a gap occurs in the monitoring data.
  • Manual migration of the AppSpeed Server virtual machine might fail with a warning if the AppSpeed Server has a virtual machine snapshot.
    Remove the snapshot of AppSpeed Server using the snapshot manager in the vCenter before performing migration.
  • After migration to ESX 3.0, the following warning might appear in the vSphere Client events log:
    The virtual machine has virtual NIC devices with non-consecutive device unit numbers. Migrating to an ESX 3 host would cause one or more of the virtual NICs to lose network connectivity.
    This has no known impact on the virtual machine.
  • In rare instances, the AppSpeed Server might incorrectly determine that traffic is being monitored by a probe other than the one that is actually performing the monitoring.
    To avoid this situation, ensure that none of the management port groups or monitored vSwitches for any of the probes are in Promiscuous mode.
  • If you upgrade vCenter Server from 4.0 to 4.0.1, you must run vCenter_settings from the console before you can connect to the user interface.

Upgrading

  • Upgrading AppSpeed from version 1.0 removes all root and local internal password settings for AppSpeed virtual appliances.
    If you specified a user password for an AppSpeed virtual appliance, you must reconfigure it following the upgrade.
  • You must manually redeploy any probes that fail to upgrade automatically.
    When you upgrade from version 1.0, all failed probes must be redeployed.
  • Probes that are manually connected to distributed vSwitches will fail to upgrade.
    Manually redeploy the probes following the deployment on the distributed vSwitches.
  • The AppSpeed VMware Tools have not been upgraded in this version and might appear as out of date.
    This has no effect on AppSpeed. AppSpeed Server does not allow upgrade of tools in this version.
  • Some VMware Tools services do not start automatically after an upgrade.
    Start the Tools services manually.

General

  • AppSpeed does not map and monitor servers or applications that use IPv6.
  • Deleting probes leaves the dedicated AppSpeed port group on the host, even after the last probe that uses the port group is removed.
    You can manually remove the port group.
  • There is no progress bar for the Delete Probe, Power On, and Power Off actions in the AppSpeed Probes Management view.
    Although it takes time for the actions to complete, it appears as if an action was performed immediately.
  • In rare cases, the virtual machine's disks of the AppSpeed Server or AppSpeed probe might become read-only. When you connect to the virtual machine console, a Read-only file system message appears.
    • If the virtual machine disks reside on a network-attached datastore, verify that the datastore is available.
    • Attempt to restart the relevant virtual machine.
      If the restart attempt is unsuccessful; in the case of an AppSpeed probe virtual machine, delete and redeploy the probe. In the case of an AppSpeed Server virtual machine, delete all the probes, reinstall the AppSpeed Server, and redeploy the probes.
  • If an AppSpeed Server virtual machine has insufficient disk space, you might be unable to connect to the AppSpeed Server user interface.
    Similarly, if an AppSpeed probe has insufficient disk space, the process on the AppSpeed probe will stop running.

    1. Delete any AppSpeed Server virtual machine snapshot.
    2. Log in to the console. If the disk is full, a message is displayed. Follow the instructions in the console.
  • The status of a probe might be Down in the Management Probe view, even though it receives traffic.
    Restart the relevant probe, either by restarting the probe's virtual machine, or by logging in to the probe's virtual machine console and restarting AppSpeed.
  • When deleting a probe, an error might appear in the Recent Tasks window and in the Probe Management window even if the probe is deleted successfully.
  • Harmless FAILED messages might indicate that the following VMware Tools failed to start: Guest memory manager, Guest VMsnet, and Guest operating system daemon.
    VMware Tools load successfully.
  • In rare cases, the AppSpeed database might become corrupted when state reset is performed on the AppSpeed Server.
    Run setup on the AppSpeed Server.
  • In rare cases, the AppSpeed Server internal database might become corrupted following a restart.
    Contact VMware technical support for assistance.
  • If an AppSpeed Server is migrated using VMotion, the AppSpeed Server might need to be restarted.
  • Virtual machines that have an identical IP address in the vCenter environment cannot be monitored by AppSpeed.
  • The following issue might arise when entering Maintenance mode on an AppSpeed Server host located on a network-attached datastore. The virtual machine does not migrate automatically when a AppSpeed Server virtual machine snapshot exists.
    Remove the snapshot and try again.
  • If a host is removed from the vCenter environment while it is in Maintenance mode, and then readded, probes that were originally deployed on the host might no longer remain on that host, and will be powered off.
    Delete the original probe and deploy a new one on the host.
  • When you enter Maintenance mode on a host that has a probe deployed, the probe is powered off and moved to another host. If the probe is manually powered on, it might result in inaccurate monitoring data.
  • If you change the vCenter Server port using the vCenter_settings command, running setup or upgrading will fail.
    Contact VMware technical support for assistance.