Brocade Communications Systems StorageWorks 4400 - Enterprise Virtual Array Manual de usuario descargar pdf (Pagina 34)

Best practices and caveats

This section includes recommendations for customers who plan to deploy VMware vCenter Site Recovery

Manager, using HP StorageWorks Continuous Access EVA. The following best practices and

recommendations should be considered by system administrators to successfully deploy Site Recovery

Manager in their virtual environment.

ESX server and Site Recovery Manager administration

 When storage is configured with VMFS on ESX 3i, you cannot dynamically resize an existing VMFS

partition. However, you can increase the volume size by using an extent to tag a new partition to an

existing VMFS datastore. HP does not recommend using VMFS extents because they tend to make the

storage replication configuration harder to manage. Instead, when running out of space on a VMFS

datastore, HP recommends creating a new LUN/VMFS data store and moving some data onto it.

Consider using Storage VMotion to avoid service interruption.

 Do not share the VMFS datastores housing I/O demanding application, such as application databases

and logs, with other virtualized applications. However, you can share a VMFS datastore to house VM

files (configuration, vmdk, and so forth).

 Configure multipathing with the FIXED policy and assign a preferred path for each LUN presented to the

ESX host. Balance LUNs across all available FC HBA to maintain an equal workload on across EVA

controllers and Fabric ISLs.

 When possible, only store protected virtual machines on the datastores that are being replicated from the

protected site to the recovery site. A careful review of the current VM locations is an important step in the

preparation for an Site Recovery Manager installation

 Protected datastores and RDM Vdisks can be replicated using one or more Continuous Access DR groups.

However, it is preferable to establish a 1:1 relationship between the Site Recovery Manager datastore

groups and Continuous Access DR groups. With straightforward mapping, the configuration is easier to

plan and maintain. The exception to the rule is a protection group, combining a single Virtual Machine

with multiple dataset spread over several virtual disks (for example, MS Exchange 2007 or SQL server

2008). In this case, VMFS or RDM disks can be spread into two or more Continuous Access DR groups to

optimize full copy and data resynchronization between source and destination arrays, using two or more

replication tunnels.

 Site Recovery Manager recovery plans executed in test mode simulate unplanned failover situation from

an application perspective. Increased disk activity may result from database application recovery activity

on the tested Virtual Machines. Size the recovery side array adequately to withstand the I/O increase or

defer test activities to take place during off peak hours.

1 2 ... 29 30 31 32 33 34 35 36 37 38

Sin comentarios

Brocade Communications Systems StorageWorks 4400 - Enterprise Virtual Array Manual de usuario Pagina 34