Red Hat Knowledgebase

Before performing guest migrations with cluster suite, all hosts involved should be updated to at least Red Hat Enterprise Linux 5 Update 1.

With a default configuration, relocating clustered virtual services during failover or using the clusvcadm command will cause the guest to be restarted rather than live migrated. The following document details the steps that should be taken on each node to enable migration:

www.redhat.com/docs/manuals/enterprise/RHEL-5-manual/Virtualization-en-US/ch19s23.html

Note that migration should only be enabled on trusted networks, as the domain virtual memory will be exchanged in raw form without encryption of the communication. In Red Hat Enterprise Linux 5 Update 1, a change to the cluster vm resource script is needed to enable live migration. In /usr/share/cluster/vm.sh in the migrate() function, change this line

err=$(xm migrate $OCF_RESKEY_name $target 2>&1 | head -1)

To this

err=$(xm migrate -l $OCF_RESKEY_name $target 2>&1 | head -1)

Now, to migrate the guest to another node use the following command:

[root@cluster2-1 ~]# clusvcadm -M vm:clustered_guest -n cluster2-2
Trying to relocate vm:clustered_guest to cluster2-2...Success
vm:clustered_guest is now running on cluster2-2

Confirm the guest is now running on the other node:

[root@cluster2-2 ~]# xm list
Name ID Mem(MiB) VCPUs State Time(s)
clustered_guest 5 511 1 -b---- 9.6
Domain-0 0 483 1 r----- 12855.5

Log in to the guest immediately to make sure it was migrated rather than restarted.