Windows HPC Server 2008 Release Notes
Applies To: Windows HPC Server 2008
These release notes address late-breaking issues and information about this release of Windows® HPC Server 2008. They contain important information that was not available when the product documentation was written. These notes will help you to install and use this product.
Known issues
The following are the known issues in this release.
SQL Server update stops all database-dependent HPC services
After SQL Server update KB948109 is installed, Windows HPC Server 2008 services do not restart. After installing this update, you need to restart the services manually. Services you need to restart are: HPC Job Scheduler Service, HPC SDM Store Service, and the HPC Management Service.
Changing the Time Zone on the head node is not supported
You cannot change the time zone on the head node after installing Windows HPC Server 2008. If you change the time zone, all the reports before that day use the old time zone, and the reports on the day of the change will be wrong. Reports after the change will have the right data for new time zone.
Upgrading from earlier versions
There is no upgrade or migration path for moving from earlier versions of Windows HPC Server 2008 to this version. You need to uninstall Microsoft® HPC Pack 2008 from the head node, reinstall this version of Windows HPC Server 2008, and then redeploy the compute nodes.
To install Windows HPC Server 2008 on a cluster that has a previous version installed
Turn off all of your compute nodes.
Export any cluster configuration that you would like to use, for example, node templates, job templates, and the compute node list as .xml files. For more information about exporting this information, see the Windows HPC Server 2008 Getting Started Guide (https://go.microsoft.com/fwlink/?LinkId=118024).
Uninstall HPC Pack 2008 from the head node.
Install the new version of HPC Pack 2008 on the head node.
Configure the head node (using the To-do List).
Redeploy the compute nodes using one of the following methods:
Turn all of your compute nodes back on, and rebuild from bare metal.
Import the list of compute nodes using a node XML file.
If your compute nodes are pre-configured, uninstall HPC Pack 2008 from the compute nodes, delete the compute node entries from the head node, and then reinstall the new version of HPC Pack 2008. Then, use the Add Node Wizard to add them back to the cluster as pre-configured compute nodes.
Manually uninstalling updates can give inconsistent diagnostic results
If you manually uninstall software updates and then run the Installed Software Updates Report, the uninstalled update will still appear as an installed update. The uninstalled updates will, however, correctly appear in the Pending Software Updates and Software Updates Required reports.
File sharing is required to shut down or reboot nodes when you enable the firewall on a private network
If you enable the firewall on the private network, you will no longer be able to shut down or reboot the nodes from HPC Cluster Manager.
To work around this issue, create a firewall exception for file sharing.
To create a firewall exception in Windows Vista
Click Start, and then click Control Panel.
Double-click Windows Firewall.
Click Allow a program through Windows firewall.
Click the Exceptions tab.
Make sure that File and Printer Sharing is checked.
To create a firewall exception in Windows XP
Click Start, and then click Control Panel.
Click Network Connections, and then click Network Properties.
On the Advanced tab, click Windows Firewall Settings.
On the Exceptions tab, make sure that File and Printer Sharing is checked.
To create a firewall exception using Clusrun
Open a command window, and type:
clusrun /readynodes netsh firewall set service type=fileandprint mode=enable profile=alland press Enter.When that command finishes, type: clusrun
/readynodes netsh advfirewall firewall set rule group="File and Printer Sharing" new enable=yesand press Enter.To verify that you can now shut down a node, type:
clusrun shutdown –r –f –t 5.
Clusrun fails when it is run from a client workstation that does not have file sharing enabled
If you do not have file sharing enabled on a client workstation, clusrun commands that are started from that workstation will fail with an error message indicating that the required named pipe cannot be opened.
To work around this issue, create a firewall exception for file sharing.
To create a firewall exception in Windows Vista
Click Start, and then click Control Panel.
Double-click Windows Firewall.
Click Allow a program through Windows firewall.
Click the Exceptions tab.
Make sure that File and Printer Sharing is checked.
To create a firewall exception in Windows XP
Click Start, and then click Control Panel.
Click Network Connections, and then click Network Properties.
On the Advanced tab, click Windows Firewall Settings.
On the Exceptions tab, make sure that File and Printer Sharing is checked.
Node Availability Report shows compute node availability for head node in a failover cluster
In the Node Availability Report, the head node is shown under the compute node availability section even though the head node cannot be configured as a compute node in a failover cluster. The head node’s compute node availability is calculated in the same way as other compute nodes (that is, it is considered to be "available" when its state is Online and its heath state is not Unreachable). If the head node is taken offline when performing node actions such as changing roles, then it is considered as unavailable and its compute node availability will be less than 100%. This issue can result in the report showing lower than actual overall compute node availability.
Updates that include a EULA cannot be installed by using node templates in HPC Cluster Manager
The Apply Updates task in node templates cannot install updates that include an End User License Agreement (EULA). This type of updates requires you to accept the EULA before you can install them. You have to install updates that include a EULA manually.