Brocade Communications Systems StorageWorks 4400 - Enterprise Virtual Array Manual de usuario descargar pdf (Pagina 3)

Solution configuration

Test scenario

To demonstrate the capability of HP disaster tolerance and disaster recovery solution, testing is

performed on a typical mid-range Microsoft Office SharePoint 2007 environment.

The SharePoint Server 2007 primary site application farm consists of the following three application

tiers:

o Web front end servers (WFEs) (3)

o Application server (index server) (1)

o SQL Server 2008 clustered nodes, which service the application back end (2)

The primary site farm is duplicated on a secondary site. For more information, see Test Environment.

During testing, HP LoadRunner software is used to simulate1,000 SharePoint users, generating a

typical collaborative workload that consists of the following:

o 40% search activity

o 20% team sites activity

o 15% document check-in check-out activity

o 10% announcement activity

o 10% event activity

o 5% My Sites activity

Approximately 0.5 TB content data is used during testing, which includes three major versions of

stored content data.

The replication example has a simulated geographical distance that does not exceed 300 miles from

the primary site (Site A) to the secondary site (Site B); or more specifically, a distance that has a

network round-trip response time of 5 ms. An Empirix Gigabit Ethernet PacketSphere XG is used as a

delay generator to simulate this distance between sites on WAN links.

When planning the solution, each tier of the SharePoint application must be considered.

WFEs

As a best practice, Microsoft recommends that there be only 1 ms delay between WFEs and the SQL

Server backend. Therefore, site-to-site failovers should include the entire application farm. For

example, if a failure occurs that causes the back end SQL server to fail over to Site B, the WFEs

should also be redirected to operate from Site B. This redirection is typically done without human

intervention when using a hardware load balancer, such as F5 BIG-IP Local Traffic Manager with

WebAccelerator. For more information, see the Microsoft Technet article, Plan for front-end web

server redundancy.

Index servers

The following challenges for availability exist for SharePoint 2007, which effects availability of

SharePoint search functionality when site failovers occur.

 Index server redundancy cannot be achieved by installing the index role on multiple servers.

 Index server redundancy cannot be achieved by clustering index servers.

To overcome the loss of an index server, Microsoft guidance states that if the business does not

require immediate search currency and availability after failover, reinstall the server and either restore

from a backup or rely on slightly stale results while SharePoint search recrawls the content. For more

1 2 3 4 5 6 7 8 ... 22 23

Sin comentarios

Brocade Communications Systems StorageWorks 4400 - Enterprise Virtual Array Manual de usuario Pagina 3