Home > Hyper-V, Virtualization > Hyper-V Virtual Machines lost Network Configuration after a failover (Updated)

Hyper-V Virtual Machines lost Network Configuration after a failover (Updated)

I started my Hyper-V failover cluster about month ago.. We are implementing HMC environment on Hyper-V failover cluster VMs. After a sudden electricity failure I started my cluster node I found that I lost my Network Configuration in all my VMs!!!!!

I had to re attach it manually..I tested that many times and every time I lost the Network Configuration.

John Huang …Microsoft Online Partner Support replied to the same problem with those recommendations.

1. Do you have identically named virtual networks on each cluster host?  Please understand that they need to be identical on both hosts otherwise the issue will happen.

2. Please manually attach the Virtual Network to the Virtual  Machine(network connection is not configured) and see if it could start  successfully.

3. Did you update the configuration of the resources after you changed the  network adapter of the already clustered VM (through the “Refresh Configuration” option in the Failover Cluster Management console)? If not, please try this to see if it can help.
The cluster resources for the VM cache information about the configuration of the VM (including information about the networks used by the VM). This information is used to configure the network on the other nodes of the cluster if the VM is moved to a different node than the VM was originally created on. If you do not refresh this information for a clustered VM after you modified the settings of the that VM through the Hyper-V manager, the Hyper-V clustering is not able to correctly configure the network on the new node and the VM will fail to restore.

Further, please install the hotfix below on each cluster nodes as this
hotifx fixed some issue for Hyper-V :
Increased functionality and virtual machine control in the Windows Server 2008 Failover Cluster Management console for the Hyper-V role
http://support.microsoft.com/kb/951308

Although I did all that tricks before his reply but nothing new..Still had the same problem.

I asked MS PPS to confirm if it is a problem with Hyper-V product or just problem with my servers.

Update: Microsoft Online Partner Support John Huang replied today


Hi Mohamed,

Thank you for your update. I know there is no broken links.

After performing a further research, I would like to confirm the following information with you:

I notice there is a known issue regarding NIC lookup. If you have multiple physical network cards and some of them are not bond to a virtual switch, the failover between cluster nodes will fail due to a known issue in Hyper-V components. Upon failover, this Hyper-V component queries available virtual switch associated with network cards and an error occurs when there is any network card that isn’t bond to virtual switch.

If so, this is known issue  and we have this reported already in our internal database and are working on to fix it.

For now, we could bind all hardware network cards to Hyper-V virtual switch to avoid this exception.

I look forward to your reply.

-regards

-John Huang

Advertisements
Categories: Hyper-V, Virtualization
  1. james
    October 15, 2008 at 4:12 pm

    are you kidding me? bind all NICs to a virtual switch is there solution!?

  2. Charles Wastell
    March 18, 2009 at 10:41 pm

    I can confirm that this is a problem, we’ve seen it on our Hyper-V failover clusters as well. It seemingly happens at random, we haven’t found a systematic reason for the problem.

  3. Mohamed Fawzi
    March 19, 2009 at 7:30 pm

    Looks like that Charles

  4. Vasil Petkov
    June 25, 2009 at 9:16 am

    You have to configure network setting on your VM BEFORE making VM cluster resource. If VM is a cluster resource you CAN NOT change Network Adapter (you have the option, but nothing will change after failover). This is not documented issue and I hope to see hotfix or something soon.

  5. Mohamed Fawzi
    June 25, 2009 at 10:06 am

    I am not sure with that.. I tried it and I can change it and just refresh the VM after the change. I think that you should refresh the VM to update the configuration.

  6. billy
    July 23, 2010 at 1:53 pm

    I had to delete the virtual networks on all nodes, then reconfigured them.

    then checked the networks in cluster manager -> al the same.

    after this life migration worked ok.

  7. Mohamed Fawzi
    July 23, 2010 at 2:12 pm

    mm.. thanks billy for this info

  8. David
    March 4, 2011 at 8:23 pm

    Hello,

    Does anyone know if this is being fixed on Windows Server 2008 R2 SP1?

    It seems I am getting this symptoms more often after I upgraded my Hyper-V server nodes to SP1 :- (

    Thanks
    Davdi

  9. Mohamed Fawzi
    March 5, 2011 at 11:59 am

    can you please send me more info to be shared with the product team

  10. TomG
    June 30, 2011 at 10:26 pm

    Billy – your solution worked perfectly for me. Kind of a pain, but got to do what we got to do. Really it boiled down to having to be 100% sure that each cluster node uses the same names too. Of course doing this does bring everything down, but it’s a good chance to document each little detail of the cluster nodes NICs and also each VM’s NICs.

    Once I deleted & re-created all NIC info, it works perfectly.

    Also Vasil’s comment was helpful too.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: