Jump to content United States-English
HP.com Home Products and Services Support and Drivers Solutions How to Buy
» Contact HP
More options
HP.com home
Insight Dynamics - VSE for ProLiant Version 4.0 Release Notes > Chapter 4 Known Issues

Logical Server Management

» 

Technical documentation

Complete book in PDF
» Feedback
Content starts here

 » Table of Contents

 » Index

Limitations

  • HP recommends that logical servers created on virtual machines and Virtual Connect server blades should be managed only by using the Logical Servers menu selections in Virtualization Manager.

    Do not use VMM, VMware ESX, VMware VirtualCenter, or Virtual Connect Enterprise Manager to perform operations on logical servers. If you use tools other than Virtualization Manager to manage logical servers, unpredictable results can occur.

    For example:

    • After you import a logical server, if you modify the host name of the operating system running on the logical server (by, for example, reconfiguring the operating system using VMware VirtualCenter), you must reimport the server or virtual machine as a new logical server.

    • If you attempt to delete a logical server using HP SIM, the logical server temporarily disappears from the HP SIM collection, but the logical server is not actually deleted and will reappear when the LSA server is restarted.

    Exceptions to this recommendation are logical servers created on virtual machines that are moved by VMware Distributed Resource Scheduler (DRS), and logical servers created on Virtual Connect blades that are moved by blade failover. In these cases, logical server management detects the new location and updates its resources to reflect the change for virtual machines and blades. You can manually resynchronize by using the Tools->Logical Servers->Synchronize... menu selection.

  • You can initiate multiple move operations at the same time, but VM Manager or Virtual Connect Enterprise Manager queues the move operations and performs them serially.

  • You can move ESX virtual machine logical servers among multiple clusters, as long as the clusters are contained in one VMware VirtualCenter.

  • You can activate an ESX virtual machine logical server only if you specified shared (SAN) or local storage as your storage configuration when you created the logical server. The only DataStore types supported are Local and Fibre Channel. Others such as iSCSI and NFS are not.

  • If you configure a SAN-based datastore (disk path) as a shared datastore, it must be attached to more than one ESX Server. When you create or modify storage for an ESX virtual machine logical server, or import a logical server, if the SAN-based datastore is not attached to multiple ESX Servers, the disk path is set to local instead of to the SAN.

  • If you are performing operations on ESX virtual machine logical servers, be sure that you have synchronized the clocks on the hosts and virtual machines. If the VM hosts are managed by VMware VirtualCenter and the hosts' clocks are not synchronized, some issues may occur while managing the virtual machines that may affect logical servers.

    Suggestion. Use clock synchronization software such as the Windows Time Service or the Network Time Protocol to synchronize the hosts' clocks.

  • After you allocate storage for a VC blade logical server (by clicking the Allocate Server WWNs button on the Create or Modify Storage screen), you cannot modify fields in the Volume Definition and Boot Volume Port Definition sections without first removing the storage, because modifying the allocated storage fields will affect the connections between the logical server and its storage.

    To modify those sections, click Remove WWNs/LUNs, enter the modified information, then click Allocate Server WWNs and save your changes.

  • The Logical Servers feature in Virtualization Manager supports 50 Virtual Connect domains (including all VC domain groups) configured on VCEM (Virtual Connect Enterprise Manager).

Major Issues

  • Occasionally, when you activate an ESX virtual machine logical server and then immediately deactivate it, you may see the following error:

    Unable to delete virtual machine

    The logical server was not deactivated and was not unregistered from VMM. If you try to move the logical server when it is in this state, you may see the following error:

    Inconsistent data on LSA inventory

    Workaround. Select the logical server, use the Delete->Delete Logical Server... menu selection, and choose Recoverable Deletion. Then manually remove the related virtual machine. You can remove the virtual machine by viewing “All Systems” in the left HP SIM pane, then selecting the virtual machine and clicking the Delete button in the popup window.

  • There is a known issue related to Onboard Administrator (OA) and Virtual Connect Manager (VCM) that could deny power for a specific bay. This may interrupt operations such as moving and activating Virtual Connect blade logical servers. Make sure that your firmware is updated to reflect the requirements for Integrated Lights-Out (iLO), OA, and VCM, and that you have the latest available update for your server blade BIOS. See “Logical Server Management Requirements” for required firmware versions.

  • If you attempt to move an ESX virtual machine logical server from one VM host to another, and the VM hosts involved in the move operation are managed by VMware VirtualCenter, occasionally the VM guest becomes unregistered on both the source and target VM hosts. Although the move operation succeeds, subsequent operations fail with the following error:

    Inconsistent data on LSA inventory. 

    This problem is related to timing and may occur intermittently. The moved VM guest may be orphaned (if the timing problem occurred during a SAN move), or may have a changed UUID (if the problem occurred during a copy move). After the move, the associated virtual machine is unusable, and will cause logical servers to be unusable until resolved.

    NOTE: HP and VMware are working together on a resolution to these issues. Please check the following website for updated VSE release information, alerts, and other important information:

    http://docs.hp.com/en/vsemgmt/index.html

    If you encounter this problem, the workaround you should use to correct it depends on the type of move you performed.

    Workaround after a SAN Move. 

    • In Virtualization Manager on the CMS, select the Logical Servers perspective and check the checkbox of the source logical server (the logical server you moved).

      • From the Report->View Logical Server Details... menu selection, obtain the name of the associated VM guest from the VM Name field.

      • From the Report->View Logical Server Details... menu selection, click the Storage tab and obtain the Storage Path of the VM guest.

    • Connect to the VirtualCenter using the Infrastructure Client.

      • Use the previously obtained VM Name to select the orphaned VM guest.

      • Remove the VM guest from the VirtualCenter inventory by right clicking the selected VM guest and selecting the Remove from Inventory menu option.

      • Select the target hypervisor.

      • Use the previously obtained Storage Path to locate the datastore, directory and configuration file of the VM guest.

      • Select the datastore where the VM files are located and open the Datastore Browser.

      • Select and open the VM guest directory.

      • Select the VM guest configuration file.

      • Re-register the VM guest by right clicking the selected configuration file and selecting the Add to Inventory menu option.

      • In the VM guest wizard:

        • Type in the previously obtained VM name.

        • Select the target hypervisor.

      • The VM will be registered on the target hypervisor.

    • In Virtualization Manager on the CMS, select the Logical Servers perspective.

      From the Tools->Logical Servers->Synchronize… menu selection, check the checkbox next to Synchronize ESX Virtual Machine Resources and click the Synchronize button.

    If this workaround is executed successfully, the logical server associated with the virtual machine will continue to be usable. The virtual machine configuration remains the same, including UUID and MAC.

    For a copy move, the workaround also depends on whether or not the virtual machine has an operating system installed.

    Workaround after a Copy Move, If the Virtual Machine Does Not Have an Operating System Installed. 

    • In Virtualization Manager on the CMS, select the Logical Servers perspective and check the checkbox of the source logical server (the logical server you moved).

      • Use the Tools->Logical Servers->Copy... menu selection to copy the logical server.

      • From the Logical Servers perspective, select the old logical server.

      • From the Report->View Logical Server Details... menu selection, obtain the name of the associated VM guest from the VM Name field.

    • Connect to VirtualCenter using the Infrastructure Client.

      • Use the previously obtained VM name to select the virtual machine.

      • Remove the virtual machine by right clicking the selected virtual machine and selecting the Remove from Disk option.

    • In Virtualization Manager on the CMS, select the Logical Servers perspective.

      • Select the old logical server.

      • From the Delete->Delete Logical Server... menu selection, click Recoverable Logical Server and Physical Server Profile or Virtual Machine Deletion to delete the old logical server.

      • In the Logical Servers perspective, select the new logical server.

      • Use the Tools->Logical Servers->Activate... menu selection to activate the logical server.

      • From the Tools->Logical Servers->Synchronize… menu selection, check the checkbox next to Synchronize ESX Virtual Machine Resources and click the Synchronize button.

    Workaround after a Copy Move, If the Virtual Machine Has an Operating System Installed. 

    • In Virtualization Manager on the CMS, select the Logical Servers perspective and check the checkbox of the source logical server (the logical server you moved).

      • From the Report->View Logical Server Details... menu selection, obtain the name of the associated VM guest from the VM Name field.

      • In the Logical Servers perspective, select the logical server.

      • From the Delete->Delete Logical Server... menu selection, click Recoverable Logical Server and Physical Server Profile or Virtual Machine Deletion to delete the logical server.

      • In the Virtual Machines perspective, select the virtual machine using the previously obtained VM name.

      • Use the Tools->Logical Servers->Import... menu selection to import the virtual machine to a new, activated logical server.

      • Select the Logical Servers perspective.

        From the Tools->Logical Servers->Synchronize… menu selection, check the checkbox next to Synchronize ESX Virtual Machine Resources and click the Synchronize button.

Minor Issues

  • If you attempt to move a logical server (by dragging and dropping the compartment in Virtualization Manager) from a server blade or virtual machine on which an operating system has been provisioned, you will not see any valid targets on which to drop the logical server unless you have previously collected utilization data about the server you want to move. To do so, run capcollect from the command line, or select Optimize->Capacity Advisor->Collect Capacity Advisor Data... from the HP SIM menu bar, and click the check box next to the name of the server blade or virtual machine on which the source logical server is active.

    Workaround. You can move a logical server by checking the checkbox of the logical server you want to move, and using the Tools->Logical Servers->Move... menu selection. Valid targets will be listed on the Assign Logical Servers to Target Hosts screen, although they will not display Capacity Advisor headroom ratings.

  • In environments with a large number of drop targets (blades or ESX hosts), drag and drop performance can degrade as the number of drop targets increases. In particular, when a logical server is selected (when a drag is initiated), it may take multiple seconds for valid drop targets to be highlighted and show their headroom ratings.

    Alternative. Moving a logical server using the Tools->Logical Servers->Move... menu selection will usually result in faster performance.

  • If you restart the HP Logical Server Automation service while logical server operations are in progress, the logical servers may become incomplete or unusable. Check the Report->Logical Server Job Status... screen to make sure that there are no running jobs before you restart the server.

  • If you restore a server blade to its default settings using a blade's BIOS, you must reset the blade from the Onboard Administrator after you activate or move a logical server. (Restoring a server blade to its default settings is an infrequent operation.) After the activate or move operation, the blade will be in an infinite loop trying to boot from a nonexistent local disk, or it may successfully boot from a local disk if one is present. If this occurs, it is necessary to reset this blade from the OA before the blade will boot from the correct boot device.

    You can detect that this situation exists in one of the following ways:

    • View the blade console via the iLO and see that the blade is in an infinite reboot loop.

    • View the blade console via the iLO and see that the blade incorrectly booted from an internal drive.

    • Detect in Virtualization Manager or HP SIM that the operating system host name for the logical server is not being discovered by HP SIM after ten minutes have elapsed from the activate or move operation. (This applies only to imported logical servers.)

    Workaround. Reset the server blade by performing the following steps.

    1. Enter the Onboard Administrator.

    2. In the left-hand pane, expand “Device Bays” and click on the affected blade bay.

    3. On the right-hand pane, click on the Virtual Devices tab and click the Reset button.

    This will cause the blade to reset and correctly accept the settings with the associated Virtual Connect profile assigned to the blade. The blade will reboot at this point to the correct SAN boot device.

  • An orphaned virtual machine is left behind when you deactivate an ESX virtual machine logical server, and occasionally when you move an ESX virtual machine logical server. The logical server remains in the Logical Servers perspective without a logical server icon.

    If you reactivate the logical server on the same hypervisor, or move the logical server to another hypervisor, these operations are performed successfully, the logical server displays the logical server icon, and the orphaned virtual machine is removed.

    However, if you unrecoverably delete the logical server after you deactivate it (without reactivating it or moving it), the orphaned virtual machine remains. These orphaned virtual machines may be unusable.

    Workaround. If you selected Delete->Delete Logical Server... to unrecoverably delete the logical server, you can use Virtual Infrastructure Client to delete the orphaned virtual machine. (This virtual machine is not marked “(orphan)” after the name in VIC.) Do not delete an orphaned virtual machine if you have recoverably deleted the logical server.

  • During an import of a server blade or virtual machine to a logical server, if the logical server database is not synchronized with the latest Virtual Connect or virtual machine elements, you may see the following error message:

    Import failed. Check that the Host is configured appropriately with VM Manager or Virtual 
    Connect Enterprise Manager on the CMS.

    Workaround. Use the Tools->Logical Servers->Synchronize... menu selection, and check the checkbox next to the type of system that you want to import (ESX Virtual Machine or Virtual Connect server blade). The synchronization may take a few minutes. When it is complete, retry the import operation. Alternatively, you can wait five minutes for the next automatic synchronization to run.

  • If you import a server blade or virtual machine to a logical server, you may see the following error message, where nn.n.nn.nn is the IP address:

    Import failed. Invalid host name specified, nn.n.nn.nn.

    Workaround. Check to make sure that the Central Management Station DNS servers are configured correctly. You can also delete the managed system host name entry in HP SIM, and then manually discover the server using its IP address (using the HP SIM menu selection Options->Discovery).

  • If you attempt to import a VM guest with multiple NIC cards to an ESX virtual machine logical server and the import fails, a NIC card may have been manually added to a VM guest using the VMware Virtual Infrastructure client after an operating system had been provisioned and its host name had been discovered by HP SIM. In this case, VMware and VMM return the IP address of the newly added NIC card as the IP address for all of the NIC cards configured on the VM guest. The return of the new IP address prevents logical server management from locating the correct VM guest from its database during the import operation.

    Workaround. Make sure that all NIC cards are added to the VM guest when the VM guest is initially created by the VMware Virtual Infrastructure client or by logical server management.

  • If you attempt to perform operations on a logical server running on an ESX Host that hosts a virtual machine using an unresponsive or inaccessible SAN datastore, you may see generic error messages such as “Unable to create virtual machine”, “Unable to delete Virtual Machine”, “Unable to migrate Virtual Machine”, and “Unable to power On/Off virtual machine”.

    This issue, although infrequent, may occur if VM files are manually removed from the host, or if the VM uses a SAN datastore that is not available because of connectivity problems or because the SAN datastore is being shut down.

    To identify this problem, connect to VMware Virtual Center or the ESX Host using the VMware Virtual Infrastructure Client. Verify that there are SAN datastores or VMs set as “Inaccessible” or VMs set as “Unknown”.

    Workaround. Remove from the host any VMs that use the inaccessible SAN and restart the VMM agent. To do so, perform the following steps:

    1. Log in to the VMware Virtual Infrastructure Client.

    2. Right click on the VM.

    3. Select Delete from Disk.

    4. In HP SIM, select the ESX Host and restart the VMM agent (using Configure->Virtual Machine Management->Restart VMM Agent...).

  • If you attempt back-to-back moves of an ESX virtual machine logical server, the second move operation will fail if you attempt it before the next automatic discovery (which runs every five minutes). If this occurs, the following error message is displayed:

    Unable to find required datastore on target virtual machine.

    Workaround. Use the Tools->Logical Servers->Synchronize... menu selection, and check the checkbox next to ESX Virtual Machine Resources. The synchronization may take a few minutes. When it is complete, retry the move operation. Alternatively, you can wait five minutes for the next automatic synchronization to run.

  • A conflict can occur if you are creating or modifying Virtual Connect blade logical server storage or modifying a storage pool entry, and you click the Allocate Server WWNs button to create a server profile when automatic discovery is running or when you or another user uses the Tools->Logical Servers->Synchronize... menu selection and checks Synchronize Virtual Connect Resources. If this happens, when you attempt to activate the logical server, the activation will fail with the following error:

    Inconsistent data on LSA inventory.

    This conflict occurs after the discover operation has started, but before the discover has called HP SIM to retrieve its current list of profiles. The amount of time for the conflict to take place increases with the number of profiles and the number of blades in enclosures managed by HP SIM.

    Workaround. Use the Tools->Logical Servers->Copy... command to copy the logical server. Use Modify->Logical Server... and select the Storage tab, then press the “Allocate Server WWNs” button to recreate the server profile. Retry the activate operation.

    In the Logical Servers perspective, select the logical server whose activation failed, and use the Delete->Delete Logical Server... menu selection. Check Recoverable Logical Server Definition Deletion.

  • If more than one user is creating or modifying Virtual Connect Blade logical server storage or storage pool entries, and these users click the Allocate Server WWNs button at the same time, the error If Boot Path is Disabled, then Fabric must be Unassigned for port #1. A Primary Boot Path needs to be specified is displayed.

    Workaround. Retry the operation by separately clicking the Allocate Server WWNs button.

  • When you activate an ESX virtual machine logical server on a VM Host, deactivate it, then reactivate it on a different VM Host, the status messages for the second activation in the Report->Logical Server Job Status... screen show that the virtual machine is created and moved instead of activated. Although these messages may appear confusing, they show the correct behavior: when reactivating a logical server on a different VM Host, the virtual machine is created on the original host, then moved to the host you selected.

  • The operating system variations displayed in the pull-down menu on the Create Logical Server or Modify Logical Server Identity screens include the Windows and Linux operating systems supported by VMware VirtualCenter 2.1. Later versions of VirtualCenter may support additional operating system variations.

Considerations

  • Occasionally, certain logical servers operations fail to completely update HP SIM. When this occurs, an active logical server compartment in the Virtualization Manager Logical Servers perspective displays Active on serialnumber (for Virtual Connect blade logical servers) or Active on VMname (for ESX virtual machine logical servers) instead of Active on hostname. If this condition remains for more than five minutes, performing a manual HP SIM discovery on the host name (using the HP SIM menu selection Options->Discovery) will usually correct the problem.

    You may want to perform a manual HP SIM discovery in the following situations:

    • After you create a logical server and provision an operating system.

    • After you reactivate or move an imported logical server.

  • Capacity Advisor headroom ratings are displayed when you assign (activate or move) logical servers to target hosts. These headroom ratings are displayed under the following conditions:

    • The Power meter is displayed only for target ESX hosts. It is not displayed for target server blades.

    • If the target host (server blade or ESX host) has not been provisioned with an operating system, the Disk Bandwidth and Network Bandwidth meters will display “0.00”, and the Power meter will display “NA”.

  • If a logical server has never been activated on the target host, the Disk Bandwidth and Network Bandwidth meters will display “0.00” for a target ESX host, and “NA” for a target server blade.

  • The following error may be displayed when you try to activate, import, or move a Virtual Connect blade logical server:

    Error retrieving HP Onboard Administrator IP address from HP SIM

    If this occurs, HP SIM might have not discovered the HP BladeSystem Onboard Administrator (OA), or the Onboard Administrator discovered by HP SIM and associated with the enclosure is not the active OA. To resolve this problem, remove both the inactive and active OA nodes from HP SIM, and manually rediscover the nodes using the HP SIM menu selection Options->Discovery.

  • Before you activate an ESX virtual machine logical server, ensure that all of your ESX Hosts (hypervisors) are configured correctly by performing the following steps.

    • Perform HP SIM Discovery: The VM Host must be discovered by HP SIM before the VMM Agent can be deployed.

      From the upper, blue menu bar, select Options->Discovery->Run Now.

    • Configure SNMP, SSH, and WBEM.

      • Select Configure->Configure or Repair Agents....

      • Deselect password setting, then set SNMP community to “public”.

      • In Step 2, enter root user and password for ESX Hosts.

      • In Step 3, select Install SSH.

    • Install VMM Agent.

      • Select Deploy->Deploy Drivers, Firmware, and Agents...->Install VMM Agent->Linux.

      • Check for failures in stdout/stderr windows.

    • Assign VMM License to VMM Agent on VM Host (all VMM installations include five demo licenses).

      • Select Deploy->License Manager.

      • Select Virtual Machine Management->Manage Licenses.

      • Select Demo->Assign Licenses.

      • Click Assign Licenses Now.

Printable version
Privacy statement Using this site means you accept its terms Feedback to webmaster
© 2007–2008 Hewlett-Packard Development Company, L.P.