2009-04-21 09:16 AM
After installing SnapDrive we can't deploy VM's using SCVMM 2008.
Prior to installing SnapDrive on our clustered hosts would use LUN's created using FilerView. With FilerView created LUN's we can attach those disks and configure as a clustered resource. Then deploy VM's via SCVMM 2008 with out a problem.
When we create a shared disk using the SnapDrive GUI it's a different story. The disk shows up as an available resource in the cluster manager MMC and is seen in VMM. However when we deploy a VM to the disk we get a hardware management error (see below). Also, if we sidestep VMM and deploy a VM to the SnapDrive created disk using the Hyper-V Manager, it works without issue.
A Hardware Management error has occurred trying to contact server host11.CWF.ORG.
(Unknown error (0x8033801a))
Check that WinRM is installed and running on server host11.CWF.ORG. For more information use the command "winrm helpmsg hresult".
has anyone seen this issue before?
2009-04-23 11:46 AM
I am suspecting some thing to do with the firewall blocking for the SCVMM server to contact the host(clusters).
Just to rule out this chance, can you try disabling firewalls in all the hosts involved in this issue.
2009-04-23 10:56 PM
2009-04-24 02:14 PM
I'm working with Andy on the system in question. Let me provide a few more details to those provided by Andy. While we can deploy to the SnapDrive created LUN's using the Hyper-V Manager .mmc, we cannot make these machines Highly Available or perform other VMM management operations on those machines. Also, even with SnapDrive installed, we can ignore SnapDrive and create new LUN's with FilerView and deploy to those LUN's using VMM. However, this leaves us unable to take snapshots or resize LUN's.
Answers to your questions:
-The firewall is completely off on all hosts and on the VMM server. I have noticed in some cases that even when the Windows 2008 Firewall is turned off, it will still block some programs/protocols. However, VMM is able to perform any function on the hosts unless the target LUN was created with SnapDrive.
-The VMM agent is currently running as Local System. I changed this to a domain service account, rebooted the host, and attempted to create a new VM. The resulting errors were the same as when running as Local System.
-I have already installed all Clustering, VMM, and 3rd party storage related hotfixes. I will list all installed hotfixes here along with the application the hotfix addresses.
-Besides this host cluster, we have 7 other non-clustered Virtual Server 2005 hosts. All hosts running Microsoft VM's and are managed with VMM.
2009-04-28 01:53 PM
I sent a trace to Microsoft and they said the New VM job was failing because it was attempting to access an invalid resource name. Apperently you cannot use a \ at the end of your resource name. Unfortunately this is the default way that SnapDrive presents drive letter mapped disks to the cluster. In our case, SnapDrive presented the LUN as S:\ . After renaming the resource in Cluster Manager and refreshing both Cluster Manager and the cluster group in VMM, the New VM job finally worked. I have not tested to see if this issue appears when using Volume Mount Points instead of drive letters.
This issue cost us a support call to Microsoft and put us nearly two weeks behind on deploying new servers. SnapDrive 6.1 is still relatively new, but it is likely that a great many more users will experience this error if the default naming is not addressed. Users can work around this by renaming it in Cluster Manager, but ideally the default naming scheme should be changed in SnapDrive.