clip_image001This is an issue that we originally talked about a while back but we're still seeing some cases come in so I thought it would be worth another mention just in case.

When using System Center Virtual Machine Manager 2008 to perform any action involving file transfers across machines such as new VM or new P2V the operation may fail with the following error message:

Error (2912)
An internal error has occurred trying to contact an agent on the %serverName%.
(Element not found (0x80070490))

Recommendation Action
Ensure that the agent is installed and running. Ensure the WS-Management service is installed and running, then restart the agent.

0x80070490 = Element not found = Certificate not found

Note: The exact error message displayed in the console is generic in nature, such as 2912 or 3154. However it will always contain the "Element no found" in the description.

This issue is caused by a problem with the Host certificate (incorrect name, IP instead of FQDN or NetBIOS) or the certificate is missing from the VMM server.

Identifying and verifying this issue requires capturing a trace using the Dbgview utility. The following article discusses this process in more detail, including where to download the Dbgview utility:

KB970066 How to collect traces in System Center Virtual Machine Manager

The following is a snippet from the stack trace while this error occurred - note where the operation is checking for the certificate on the VMM server. Another key is the BitDeployer.Copy function.

00006921 302.15213013 [1972] at Microsoft.VirtualManager.Engine.Deployment.FileDeploymentServerJobFactory.GetPeerCertificate(WsmanAPIWrapper clientWrapper)
00006922 302.15213013 [1972] at Microsoft.VirtualManager.Engine.Deployment.ImageDeploymentServerJobFactory.CreateJob(WsmanAPIWrapper serverWrapper, WsmanAPIWrapper clientWrapper, DeploymentFile file, Boolean useHTTPS)
00006923 302.15213013 [1972] at Microsoft.VirtualManager.Engine.Deployment.BitDeployer.Copy()
00006924 302.15213013 [1972] at Microsoft.VirtualManager.Engine.Deployment.DeploySubtask.RunSubtask()
00006948 302.15328979 [1972] *** Carmine error was: HostAgentFail (2912); HR: 0x80070490

The solution is to remove the managed host from the VMM server and also delete any residual certificates from the host on the VMM server, and then re-add the host:

  1. On the SCVMM server, remove the managed host from the console. The steps on how to remove a managed host are outlined in the following TechNet article:
    http://technet.microsoft.com/en-us/library/cc956121.aspx (http://technet.microsoft.com/en-us/library/cc956121.aspx)
  2. Now we need to locate and delete any certificates for the Host computer.
  3. Open the Certificate console on the SCVMM server.
    a. Open a new mmc and add the certificates snap-in.
    b. Select the option of 'computer account' and 'local computer'.
    c. Select Finish and Ok to load the snap-in.
  4. The certificates for the Host computer can be in any of the following locations.
    a. Personal Certificates.
    b. Trusted People (if the host is W2K8).
    c. Trusted Root Authorities (If the host is W2K3).
  5. In each store, expand the Friendly Name field and locate the certificate[s] for the Host server that have a Friendly Name starting with 'SCVMM_CERTIFICATE_KEY_CONTAINER' followed by either the FQDN / IP address / NetBIOS name of the Host server and delete them.
  6. Re-add the host in SCVMM which recreates the certificates as needed.

More Information:

SCVMM uses BITS to transfer payload between SCVMM managed computers. These data transfers are encrypted by using a self-signed certificate generated at the time a host machine is added to VMM. If these certificates are missing or corrupted from the VMM server or managed computers, the payload deployment job can fail. Deleting the certificates and re-adding the host will cause the certificates to regenerate.

This same information is documented in our Knowledge base at http://support.microsoft.com/kb/971264.

J.C. Hornbeck | System Center Knowledge Engineer