HPC Pack 2016 installation fails when creating Service Fabric application RRS feed

  • Question

  • I try to install HPC Pack 2016 Update 1 on a Windows Server 2016 Version 10.0.14393 and the installation repeatedly fails when installing the service fabric. As the Powershell windows are instantly closed on error, it is very difficult reading anything, so I am not sure whether there is additional information besides the stuff logged to C:\Windows\Temp\HPCSetupLogs\HPCSetupLogs-...\ServiceFabricCluster.err:

    Message        : Method not found: 'System.String 
    Data           : {}
    InnerException : 
    TargetSite     : Void CopyApplicationPackage(System.String, System.String, System.String)
    StackTrace     :    at Microsoft.ServiceFabric.Powershell.ApplicationCmdletBase.CopyApplicationPackage(String 
                     imageStoreConnectionString, String applicationPackagePath, String relativeDestinationPath)
                        at System.Management.Automation.CommandProcessor.ProcessRecord()
    HelpLink       : 
    Source         : Microsoft.ServiceFabric.Powershell
    HResult        : -2146233069

    I see no way to obtain any further information about the error. As I do not really know SF and am only interested in the HPC pack, I also have no clue how to debug this issue. The SF status is like:

    AggregatedHealthState   : Warning
    UnhealthyEvaluations    :
                              Unhealthy event: SourceId='System.UpgradeOrchestrationService',
                              Property='ClusterVersionSupport', HealthState='Warning', ConsiderWarningAsError=false.
    NodeHealthStates        :
                              NodeName              : vesta2
                              AggregatedHealthState : Ok
                              NodeName              : vesta1
                              AggregatedHealthState : Ok
                              NodeName              : vesta3
                              AggregatedHealthState : Ok
    ApplicationHealthStates :
                              ApplicationName       : fabric:/System
                              AggregatedHealthState : Ok
    HealthEvents            :
                              SourceId              : System.UpgradeOrchestrationService
                              Property              : ClusterVersionSupport
                              HealthState           : Warning
                              SequenceNumber        : 131656822793242186
                              SentAt                : 16.03.2018 13:57:59
                              ReceivedAt            : 16.03.2018 13:58:02
                              TTL                   : 2.00:00:00
                              Description           : The current cluster version support ends 30.03.2018
                              00:00:00. Please view available upgrades using Get-ServiceFabricRegisteredClusterCodeVersion
                              and upgrade using Start-ServiceFabricClusterUpgrade.
                              RemoveWhenExpired     : True
                              IsExpired             : False
                              Transitions           : Ok->Warning = 16.03.2018 13:58:02, LastError = 01.01.0001 00:00:00

    Any suggestions?

    Thanks in advance,

    Friday, March 16, 2018 2:06 PM

All replies

  • Hi Christoph,

    Had you installed Service Fabric related module before (for example Service Fabric SDK, service fabric PowerShell)? 

    Could you try to reboot the head node and re-install?

    Monday, March 19, 2018 9:12 AM
  • Hi Sunbin,

    for me having absolutely no experience with SF, I just ran the installer and let it do everything. The target machine was a newly installed Server 2016 (with GUI). I have also repeated the procedure several times (removing SF via the cleanup scripts and all of HPC Pack via control panel), but the problem could be reproduced every time.

    As I had to get the thing running until Monday, I have installed HPC Pack 2015 (w/o Update 1), which worked after uninstalling everything of Update 1. I was very reluctant doing so because the update process looks ... complicated at best. However, the next step is bringing the head nodes to Update 1 via the upgrade scripts.

    Best regards,

    Tuesday, March 20, 2018 9:04 AM