HPC Pack 2012 R2 Update 3 Fix for SOA, Management and Azure burst issues

This update fixes some known SOA, Management and Azure burst issues of HPC Pack 2012 R2 Update 3:

SOA improvements
- Fixed the error "Authentication failed. Make sure you have the permission to the SOA diagnostics service." when SOA message level logs are viewed by administrators from a trusted domain other than the domain of the job owner.
- Fixed the issue when setting serviceOperationTimeout to a small value for a SOA service, that the timeout exception for each request from the service cannot be passed through to the client.
- Fixed the issue that jobs could be stuck in the queued state due to a SOA job cannot be preempted because of a syncing issue between scheduler and the broker when there are a lot of responses waiting to be flushed into MSMQ and only a few processing requests.
- Increased the session creation throughput when using a shared session pool.
- Provided the binding parameter to override the default binding when creating SOA sessions. To use the new Session API, please download the QFE version of the SDK.

Management and Azure burst improvements
- Fixed the issue in HPC Cluster Manager, in Job view, when user selects one “Admin Jobs”, then clicks “Job Details” property tab, it is unable to show job details.
- Fixed a backward compatibility issue for the GUI when connecting to HPC Pack 2012 R2 Update 2 (and previous version) clusters.
- Provided two new configurations by registry keys to alter the node state and the default 60 minute operation timeout for Azure burst deployments.
1. HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\HPC Name:Microsoft.Hpc.Azure.HpcSyncFailureEnable Type:REG_DWORD Data:1
By default, HpcSync results don't affect the final Azure node deployment state. After configuring this registry key with DWORD value 1, when HpcSync fails to sync all the packages on a node, the node will be marked in the Error state and this won't block completion of the whole deployment operation. The HpcSync will continue to retry downloading the packages on the nodes with a default 5 minute interval and once this succeeds, the Azure node will automatically reach the OK state. The deployment operation logs will show the related message for any Error nodes because of the HpcSync failure. This registry key shouldn't be set if there is no user package being uploaded through hpcpack.exe otherwise the provision will fail as hpcsync does not find any packages
2. HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\HPC Name:Microsoft.Hpc.Azure.DeploymentOperationTimeoutInMinutes Type:REG_DWORD Data:60
By default, the Azure burst deployment operation has a timeout of 60 minutes. With this registry key, the deployment timeout can be customized according to the requirement. For example, change it to 30 minutes if a shorter operation timeout is preferred.

Scheduler fixes
- fix issue that task may incorrectly created with dependency while the job is already running
- fix performance issue for jobs with large number of tasks with dependencies

Supported Operating Systems

Windows 10, Windows 7, Windows 8, Windows 8.1, Windows Server 2008 R2, Windows Server 2012, Windows Server 2012 R2

HPC Pack 2012 R2 Update 3 (build 4.5.5079.0) with QFE KB3134307 (build 4.5.5094.0) installed.

This update needs to be run on all head nodes, broker nodes, workstation nodes, compute nodes and clients

Before applying the update, please check if HPC Pack 2012 R2 Update 3 with KB3134307 is installed. The version number (in HPC Cluster Manager, click Help>About) should be 4.5.5094.0. Please take all nodes offline and ensure all active jobs finished or canceled. If there are Azure nodes, please make sure they are stopped before applying this patch. After all active operations on the cluster have stopped, please back up the head node (or head nodes) and all HPC databases by using a backup method of your choice.

Applying the update

To start the download, click the Download button next to the appropriate file (KB3147178-x64.exe for the 64-bit version, KB3147178-x86.exe for the 32-bit version) and then:

1. Click Save to copy the download to your computer.
2. Close any open HPC Cluster Manager or HPC Job Manager windows.
Note: Any open instances of HPC Cluster Manager or HPC Job Manager may unexpectedly quit or show an error message during the update process if left open. This does not affect installation of the update.
3. Run the download on the head node using an administrator account, and reboot the head node.
Note: If you have high availability head nodes, run the fix on the active node, move the node to passive, and after failover occurs run the fix on the new active node. Do this for all head nodes in the cluster.
4. Log on interactively, or use clusrun to deploy the fix to the compute nodes, broker nodes, unmanaged server nodes, and workstation nodes.
To use clusrun to patch the QFE on the compute nodes, broker nodes, unmanaged server nodes, and workstation nodes:
a. Copy the appropriate version of the update to a shared folder such as \\headnodename\HPCUpdates .
b. Open an elevated command prompt window and type the appropriate clusrun command for the operating system of the patch, e.g.:
clusrun /nodegroup:ComputeNodes \\headnodname\HPCUpdates\KB3147178-x64.exe -unattend -SystemReboot
clusrun /nodegroup:BrokerNodes \\headnodname\HPCUpdates\KB3147178-x64.exe -unattend -SystemReboot
Note: HPC Pack updates, other than Service Packs, do not get automatically applied when you add a new node to the cluster or re-image an existing node. You must either manually/clusrun apply the update after adding or reimaging a node or modify your node template to include a line to install the appropriate updates from a file share on your head node.
Note: If the cluster administrator doesn’t have administrative privileges on workstation nodes and unmanaged server node, the clusrun utility may not be able to apply the update. In these cases the update should be performed by the administrator of the workstation and unmanaged servers.
c .To update workstation nodes and unmanaged server nodes you may need to reboot.

5. To update computers that run HPC Pack client applications apply the following actions:
a. Stop any HPC client applications including HPC Job Manager and HPC Cluster Manager
b. Run the update executable
c. Reboot your client computer

6. You can run get-hpcpatchstatus.ps1 on headnode/computenodes/client under %CCP_HOME%bin to check the patch status. And the client version and server version will be 4.05.5102.0