HP EVA Cluster Extension Software Administrator Guide (5697-2445, July 2013)

An alternate method is to manually move all resources to another node in the cluster before stopping
the cluster service. After all resources are successfully moved to another node, stop and then restart
the cluster service; then, manually move back all resources.
Logs
If the EVA Cluster Extension log files need to be cleared and reset (for example, to reduce disk
space usage), you can delete the files. EVA Cluster Extension automatically creates new log files.
NOTE: When VM management functionality is used on a Windows Server 2012 system, the
logs for this feature is created in the same location. The name of the file is VMClxMscs.log.
TIP: Archive the log files before deleting them.
Hyper-V Live Migration log entries
In the EVA Cluster Extension log file (clxevamscs.log), live migration messages include the
prefix CLX_LM to help you differentiate live migration issues from EVA Cluster Extension log
messages.
For example:
[10/12/09 20:13:02][2136][CLX_LM: CLXVMDISK04-App01][INFO] CLX detected
that Live Migration for VM "Virtual Machine VM04" has begun.
[10/12/09 20:13:02][2136][CLX_LM: CLXVMDISK04-App01][INFO] CLX started
gathering VM "Virtual Machine VM04" specific storage information.
Virtual Machine Management hosted on a Cluster Shared Volume on Windows Server
2012
Cluster Shared Volumes on Windows Server 2012 support virtualized workloads to create virtual
machines on Cluster Shared Volumes in the Failover Cluster Manager. These VMs are highly
available as they are managed by the Failover Cluster.
CSV disk resource can be owned by one cluster node whereas, highly available VMs that are
created on the CSV disk can either be owned by the same cluster node or by a different cluster
node. The node that owns the CSV disk resource is known as the co-ordinator node. In a multi-site
storage array replicated environment, clustering of nodes is distributed across the datacenter. For
non-CSV disks, the application role such as VM role contains the application resource, disk resource,
and the CLX resource together. That is, in whichever cluster node the disk resource is online, the
application resource such as VM resource will also be online. Therefore, the entire application
role containing the application resource, disk resource, and the CLX resource together will be
owned by a cluster node. The CSV feature ensures that the CSV disk can be owned by any cluster
node and highly available VMs that are created on the CSV disk running either on the same cluster
node, or on a cluster node different from that of the CSV.
If the highly available VMs are running on a cluster node that is in a different datacenter, the VMs
run in the redirected-access mode. This is because, the replicated disk in the datacenter where
VMs are running is presented to the cluster node as read-only on the storage array. This indicates
that, in the CSV disk, the I/O operations of the VMs is carried over the IP network to the coordinator
node in the other datacenter. The coordinator node ensures that the I/O is committed to the disk.
Because of this reason, VMs might incur a performance penalty especially if the distance between
the datacenter is long. The CLX VM management feature is applicable in such a scenario.
The CLX for VM management ensures that the highly available VMs and the co-ordinated node of
the CSV disk are always running in the cluster nodes of the same datacenter.
Administration 45