White Papers

BP1013 Best Practices for Enhancing Microsoft Exchange Server 2010 Data Protection and Availability 46
VSS timeout
Under rare circumstances of extreme load, the VSS service can time out when a requester
makes an inquiry to create a shadow copy. Follow Microsoft resolution advice to correct this
problem. Evaluate an increase of the VSS timeout value in the registry configuration editor if
required.
General network best practices
Use separate network infrastructures for the isolation of the LAN traffic from the SAN traffic
(iSCSI).
Use redundant elements (switches, ISLs) to provide a reliable network infrastructure
Enable flow control for the switch ports where the PS Series arrays are connected
Enable jumbo frames (large MTU) for the switch ports where the PS Series arrays are
connected
Disable spanning tree for the switch ports where the PS Series arrays are connected, and
enable Portfast instead
Evaluate jumbo frames (large MTU) for the LAN network when appropriate (limited by type of
devices the traffic traverses)
General storage best practices
Dedicate separate pools for databases connected to nodes in a DAG. Do not share the same
disks for active and passive copies of an Exchange mailbox database.
Choose the appropriate RAID level.
o RAID-50 offers a good combination of performances and capacity requirements for
Exchange mailbox database deployments
o RAID-10 offers greater write data operations at a lower capacity levels
o RAID-5/6 offer good performances and greater capacity. Supported by Microsoft, and
requires careful validation if considered for deployment.
Distribute the controllers network port connections appropriately to each network switch
Use MPIO DSM provided by EqualLogic HIT Kit
General ESX and Virtual Machines best practices
Use separate virtual switches for Virtual Machines with network traffic and iSCSI storage traffic
Use Port Group and VLAN tagging to logically segregate different kinds of LAN traffic
Enable jumbo frames (large MTU) for the iSCSI virtual switch
Evaluate jumbo frames (large MTU) for the Virtual Machines virtual switch as well
Use at least two physical network adapters for each virtual switch to achieve redundancy
Use a guest OS iSCSI initiator when using ASM/ME to allow client integration with the SAN
Use performance optimized network adapters of the type VMXNET3 for guest iSCSI
connections (VMware tools required)
Enable TSO and LRO on the guest network adapters for iSCSI storage traffic
Do not allocate more virtual CPU than is really required in the virtual machines
Do not overcommit memory on ESX hosts. Exchange virtual machines have a memory
intensive workload.