How To Fix Hyper-V “Migration Attempt Failed” When Live Migrating
Posted by Matthew Hodgkins in Hyper-V Mar 3rd, 2010 | 14 responsesIf you are running Windows Server 2008 R2 with Hyper-V and have a SAN available, Hyper-V Live migration is a great way to make your Virtual Machines highly available. There is nothing like being able to migrate your VM’s to another node in the cluster and shutdown a host to perform maintenance – I still get excited every time I live migrate a machine!
In some cases though, your Live Migration attempt may fail. Unfortunately the Failover Cluster Manager nor the event viewer sheds much light on the reason in most cases.
To troubleshoot a problem with virtual no error message is difficult, and my process was to do a Google on the event log error and try my luck with the different forum posts from the TechNet Forums and other such sites. For this reason I decided to make a checklist to go through to try and work out the reason behind the failed live migration. Good luck!
The “None Of My Hyper-V Servers Are Live Migrating” Checklist
 |
In the Failover Cluster Manager do your clustered networks have a status of Up? |
|
Do you have enough RAM free on the server you are trying to Live Migrate to? The best way to check this is System Center Virtual Machine Manager. If you do not have this, checking task manager on the destination server will show you how much RAM you have free. |
|
In the Failover Cluster Manager do your Clustered Shared Volume(s) has a status of Online? |
|
In the Failover Cluster Manager do your clustered Storage volumes have a status of Online? |
|
Are you able to ping the node you are trying to Live Migrate to and visa-versa? |
|
Did you create the Hyper-V network adapters BEFORE creating a cluster? If not you should destroy your cluster and start again reading the Microsoft guide Using Hyper-V and Failover Clustering |
|
Are the names of your Hyper-V Virtual Adapters all the same? |
The Hyper-V “Migration Attempt Failed but other VM’s migrate fine” Checklist
NOTE: This post is meant for people that are having a SINGLE Virtual Machine failing to Live Migrate for example: VM1 and VM2 are hosted on NODE1. VM1 is successfully Live Migrated to NODE2 but VM2 says “Migration Attempt Failed”.
|
Make sure no CD/DVD or image file is mounted in the Virtual Machine you are trying to migrate. |
|
Reset the permission on the folder on the SAN that your Virtual Machine resides in to Everyone has full permission (this is not a security issue as the SAN network should be on a separate IP to your workstations and only Hyper-V administrators should have permission to access the cluster and its nodes) |
|
This is a weird one: Do you have any notes in the Name field under the Management section of the settings for the Virtual Machine? If so remove it and try Live Migration again. I personally have had this issue and removing the notes in here (that I never entered) worked. |
|
If you have Anti-Virus software installed on one of your nodes, ensure it is set not to scan the C:\ClusterStorage folder |
Hopefully using these checklists assists you in fixing your Hyper-V Live Migration problems. I will endeavour to update this list as I discover more reasons that Live Migration might fail.
Nice checklist – one more to add – before attempting a live migrate make sure you have sufficient free memory on the target Hyper-V host, otherwise the live migrate will fail with an unhelpful “failed” message. There is currently no simple way to see how much memory has been pre-allocated to virtual servers on a Hyper-V host. Error log entries observed will include:
Could not receive data from the remote node: An established connection was aborted by the software in your host machine. (0×00002745)
Migration failed at source for cluster resource ‘Virtual Machine {server}’ with events ”Virtual Machine {server}’ live migration did not succeed at the source.
@Simon – thanks for that – I have updated the post according to your suggestion. I’m glad you liked the list
I had the problem with one server not live migrating…
I removed the incluster variant name…and voila..the server jumped over to the second node..
@yay – glad the tip worked
I had to disable my antivirus program to use cluster shared volumes. I would not let me start a VM from a machine if it wasen’t the owner of the CSV. Disabling the AV solved that issue.
@Thomas – Thanks for the tip – I have updated the blog post accordingly.
Thanks for a great post but no luck here
.
One machine not live migrating, no cd, everyone has full control, note removed no AV.
Any other ideas?
A Reboot solved the issue…
@Tirano – Funny how many times reboot fixes issues, but how many times we forget to do it
Same thing happened to me today hehe
Fantastic tutorial. Thanks. It saved me a heck of a lot of grief! Seriously, thanks!!
Thanks! The Virtual Adapter names got me. This is a very helpful list for errors with no additional information.
Matthew, just wanted to add that you don’t have to destroy and re-create the cluster if the Virtual Networks are created later on. You can evict the node you have made changes and re-join while proper networks are present.
Also, if a particular VM is having the issue, exporting that VM and re-importing it to the node with the proper cluster networks and adding it as a cluster resource also fixes the problem.
Great checklist.
Unfortunately for me nothing helped, ’til i found the solution on the microsoft site
http://support.microsoft.com/kb/2475761
For me it was the issue of having an additional network adapter created afterwards.
This i was able to solve by refreshing the virtual machine configuration, see points 9. to 12. on the upper microsoft site.
Got a addition to the list that may be helpful.
If one of the nodes has an ip address in another subnet which the other nodes dont have, this may also cause problems.