locked
Live Migration Attempt Failed (Status) but VM is working fine and migrated successfully RRS feed

 > 

  • Question

  • Question
    Sign in to vote
    0
    Sign in to vote

    We have a Five node Hyper-V failover cluster. when i m doing live migration its give me error "Online(Migration attempet failed)" although VM migrated successfully to th other node and working fine.when i check the resource critical events it say:

    EventID: 21502

    " Virtual Machine SRVVMM01" live migration did not succeed at the source.Live Migrtaion did not succeed. The operation timed out.

    her are some configutaions of our Hyper-V failover cluster:

    - we are not using CSV

    -each node has 4 NIC's 1 for heartbeat, 1 for Live migration and 2 teamed NIC's for VM's

    Regards,

    Nouman

     

    Tuesday, January 25, 2011 9:34 AM

Answers

  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi.

    Had a similar issue on a 3 node Hyper-V Cluster. Turns out that the server copies every block in memory accross till the destination host has exactly the same blocks in memory as the source host for the specified VM, and only then the migration is officially "complete"

    Problems may occur if your "Live Migration" network is over utilized or insuficcient in capacity - and even though the migration takes place and seems to work fine -  not all blocks of memory has been written accross, and it pop's up with the above error.

    Im may also be that theye is not enough "quiet" time on the server and the memory blocks changes too quickly for the destination host to confirm the migration as complete.

    Some suggestions:

    1. You can increase the amount of memory in the VM that's migrating - for some reason this worked for me.

    2. Try increasing your Page File sizes on the VM's, this should give you more "quiet" time on actual RAM for the VM, and in theory the Live migration should complete.

    Friday, January 28, 2011 9:11 AM

All replies

  • Question
    Sign in to vote
    0
    Sign in to vote

    Drop VMM from the picture here and test LM functionality from the Failover Cluster UI.  Does it work there?  All VMM does with LM is call into Cluster to perform this action.  So start your troubleshooting in Failover Cluster.

    ____________________________________________________________________________________________________________
    Best Regards, Mike Briggs [MSFT] – Posting is provided "AS IS" with no warranties, and confers no rights

    Tuesday, January 25, 2011 4:12 PM
  • Question
    Sign in to vote
    0
    Sign in to vote
    we are already using the Failover cluster UI to do the Live migration ......  but the result is same....
    Wednesday, January 26, 2011 8:53 AM
  • Question
    Sign in to vote
    0
    Sign in to vote

    So with VMM out of the picture the same error occurs using FCMgr, then that's where I'd recommend beginning your troubleshooting.  Look in the cluster event logs for additional information as to why its failing.  What are those events relating to the failed migration?  Why does cluster say the migration failed? 

    Also, run a full cluster validation against all nodes and reply back with its results.  Pay particular attention to any warnings and/or errors.

     

    ____________________________________________________________________________________________________________
    Best Regards, Mike Briggs [MSFT] – Posting is provided "AS IS" with no warranties, and confers no rights

    Wednesday, January 26, 2011 5:46 PM
  • Question
    Sign in to vote
    0
    Sign in to vote

    Hi.

    Had a similar issue on a 3 node Hyper-V Cluster. Turns out that the server copies every block in memory accross till the destination host has exactly the same blocks in memory as the source host for the specified VM, and only then the migration is officially "complete"

    Problems may occur if your "Live Migration" network is over utilized or insuficcient in capacity - and even though the migration takes place and seems to work fine -  not all blocks of memory has been written accross, and it pop's up with the above error.

    Im may also be that theye is not enough "quiet" time on the server and the memory blocks changes too quickly for the destination host to confirm the migration as complete.

    Some suggestions:

    1. You can increase the amount of memory in the VM that's migrating - for some reason this worked for me.

    2. Try increasing your Page File sizes on the VM's, this should give you more "quiet" time on actual RAM for the VM, and in theory the Live migration should complete.

    Friday, January 28, 2011 9:11 AM
  • Question
    Sign in to vote
    0
    Sign in to vote
    Solution for this and all other Hyper-V Clustering, CSV and Live Migration related problems, are documented here:
    http://zahirshahblog.com/2012/01/19/hyper-v-live-migrati-on-is-not-working-the-disk-structure-i-s-corrupted-and-unre-adable-cluster-res-ource-virtual-machi-ne-vm-in-clustered-service-or-applicat-ion-vm-failed-vi-rtual-machine/ 
    Zahir Hussain Shah | MVP - Exchange Server | Senior Infrastructure Consultant - Messaging | My blog: http://zahirshahblog.com ]If my answer fixes your problem, mark them as answer, so it will help others to find a solution]
    Thursday, January 19, 2012 9:57 AM