beautypg.com

Cluster workloads – HP Storage Mirroring Software User Manual

Page 625

background image

Recommended optimizations

Page 624 of 677

Cluster workloads

You should carefully review Microsoft documentation and resources for properly
configuring your cluster before implementing Storage Mirroring Recover on a cluster.
The Microsoft TechNet articles

Failover Clusters

and

Installing and Upgrading Cluster

Nodes

are two resources you can start with. There are many other resources available

on the

Microsoft TechNet web site

.

l

Double-Take Source Connection resource—In a clustered environment, MSCS
handles failover between nodes of the cluster. By using the Double-Take Source
Connection resource, you can further protect your data by replicating the cluster
data to a target. The Double-Take Source Connection resource automatically
starts mirroring and replication after a group changes nodes on the source. In the
event the cluster fails, your cluster data will be available on the target.

l

Target cluster—If your target is also a cluster, connect the Double-Take Source
Connection resource to a virtual IP address on the target.

l

Auto-reconnect—Storage Mirroring Recover mirroring should be controlled by the
Double-Take Source Connection resource. Therefore, you must disable the default
auto-reconnect process for each node in the cluster. This option is available
through the

Server Properties Setup tab

. Disable the option Automatically

Reconnect During Source Initialization.

l

Storage Mirroring Recover disk queue—Ensure that the disk queue is not on a
Physical Disk resource. See the

installation optimizations

and

Queuing data

for

more details on the disk queue.

l

Exchange or SQL protection—If your source is a cluster running Microsoft
Exchange or SQL, use the Storage Mirroring Recover Application Manager,
instead of trying to use the Double-Take Source Connection resource. The
Application Manager does support clusters for Exchange and SQL.

l

Cluster service account—For Windows 2003 clusters, use the same cluster
service account on source and target clusters.