When installing the Cloud Services Process Pack SP1 version, I ran into a few issues that are worth noting. On the Orchestrator Server

  1. If you don’t have an OrchestratorUsersGroup on the local machine because you selected a domain account during the install. You will need to create one this was covered by a blog post you can find here.
  2. Add your service account for Orchestrator and the account running setup if they are different to the OrchestratorUsersGroup.
  3. SQL Version Requirement, I didn’t find any documented requirements searching the web and the admin guide quickly. On the configuration screen below I was getting an error “SQL Server 2008 was not found” My SQL Server was running SQL 2008 R2 SP2. After upgrading to SQL 2008 R2 SP2 CU5 I was able to proceed past the error.

     

    I am now stuck at the step of importing the runbooks

    One thing to check is that your powershell scripts are unblocked you can find details on that here, but that didn’t allow me to proceed any further in the setup.

    Update: I was able to finish the install

    The last change I made that allowed the setup wizard to finish was to open the runbook designer on the server, create a new runbook check it out, then delete it.

    Few other items to note that I had recently done in troubleshooting was added the logged on user to the OrchestartorSystemGroup on the local machine and added the service account to the OrchestratorUserGroup on the local machine.

    Here is the Technet forum post that provided the solution to the strange problem.

 

Today I figured out an issue that has been holding up moving forward with my deployment of Hyper V. I didn’t find much on the web when I first ran into this issue so I hope this helps someone else in the future.

Background

Building a 5 node Windows 2012 Hyper V with a CSV the storage is presented from a HP EVA 6000. Initially the build was fine until I started testing failover testing. On one of the node reboots storage failed to come back online. This led me to do some additional testing on the Cluster Shared Volume, found that I could not change the owner of the volume.

System
Provider
[ Name] Microsoft-Windows-FailoverClustering
[ Guid] {BAF908EA-3421-4CA9-9B84-6689B8C6F85F}
EventID 1069
Version 2
Level 2
Task 3
Opcode 0
Keywords 0x8000000000000000
TimeCreated
[ SystemTime] 2013-03-05T06:02:22.956203200Z

 

EventRecordID 30959
Correlation
Execution
[ ProcessID] 6876
[ ThreadID] 6252
Channel System
Computer HV04.domain.local
Security
[ UserID] S-1-5-18

 

EventData
ResourceName Cluster Disk 1
ResourceGroup ff5467c6-38dd-42f0-b7a3-6cbcde5d0d32
ResTypeDll Physical Disk
ApplicationErrorCode Element not found.
ErrorCause 0x490

After a lot of double checking everything and updating the firmware on the EVA and all of the drivers on the server. The problem still persisted but I was able to verify at this point everything was running with drivers for Windows 2012. So this either had to be a bug or a setting that was incorrect. It turned out to be a host persona setting in the EVA.


The Operating system type was set on windows on some of the host in the cluster, after changing the setting to Microsoft Windows 2008, I was able to change the owner of the CSV without any problems.

Incase anyone runs into the issue there is a bug in VMM SP 1 Beta that doesn’t support Hyper V Virtual Switches to be bound to  teamed NICs in Server 2012.

symptoms I experienced were: Logical networks did not populate in VMM correctly, Virtual switches didn’t show up in VMM, when you could get the virtual switches to show up in VMM  and removed them to rebuild them they destroy the NIC team in addition to the virtual switches.

The solution to this is to bind the virtual switches to a single NIC then everything works.