Adaptec 2074700-R - ICP Vortex GDT8514RZ RAID Controller, Руководство пользователя

Страница: 198 / 214

198 Software Installation and User’s Guide
Optional Features
The whole process (take over of tasks for the other Nodes) is classed as a Failover by
Microsoft.
Figure A-2. Microsoft Clustering
Furthermore, applications can operate on various Nodes, ensuring that all available
resources are used to the optimum. An application can not run on both Nodes at the same
time (Load Balancing). Microsoft therefore defined this as Static Load Balancing, for
example, a static distribution of resources. Dynamic load balancing is not possible at the
moment. A speed advantage is only possible if, for example, two different applications are
running in parallel on different Nodes of a cluster server (for example, a Web Server on
Node 1 and an SQL database server on Node 2).
Both Nodes can communicate with each other via the normal network (User LAN). It is
possible to link the Nodes with an additional network (Private Network).
Microsoft urges the user to use two Interconnects (1 Public Network as User LAN and a
Private Network for the Nodes) for OEMs with original MSCS equipment. The diagram
on the previous page shows this kind of configuration.
A software RAID solution (for example, the NT build-in) is not supported. Therefore, data
that is on the hard disks of the shared storage bus must be protected by hardware RAID
controllers. Microsoft recommends using suitable hardware RAID controllers for
protecting data. It is important for these controllers that the RAID configuration data is
stored on the disks and not in an NVRAM. This guarantees that even if a controller fails,
or after a controller has been exchanged, the RAID configuration still remains intact.