Recommendations: Storage (FAST Search Server 2010 for SharePoint)

FAST Search Server 2010

Applies to: FAST Search Server 2010

Topic Last Modified: 2011-03-15

This article describes storage recommendations for Microsoft FAST Search Server 2010 for SharePoint. For a complete recommendations overview, refer to Performance and capacity recommendations (FAST Search Server 2010 for SharePoint).

In general, we recommend high performance storage. High performance storage provides improved query performance, especially during concurrent content crawls, and reduces or even eliminates the performance driven need to run search on dedicated rows.

FAST Search Server 2010 for SharePoint can use SAN storage instead of local disks if necessary for operational reasons. Testing has shown that a sufficiently powerful SAN will not be a bottleneck. Although the actual workload is scenario dependent, the following parameters can be used as an initial estimation of the required SAN resources for a server in the FAST Search Server 2010 for SharePoint farm:

  • 2000 – 3000 I/O operations per second (IOPS)

  • 50 – 100 kB average block size

  • Less than 10 ms average read latency

For example, for a farm setup with 7 servers, the SAN must be able to serve 15.000 – 20.000 IOPS to the FAST Search Server 2010 for SharePoint farm regardless of any other traffic the same storage system may serve.

FAST Search Server 2010 for SharePoint can take advantage of the increased IO performance of solid state disks. Deployments with low content volume, high QPS and regular spindles will typically require many disks, whereas testing has shown that if you are using solid state disks, it can be sufficient with only a couple. This allows for using blade servers that have local disks on the blade itself, and still having sufficient IO performance.

SSDs make it possible to have high query performance without a dedicated search row, and let you cut the server count. But be aware that removing the search row also removes redundancy and reduces availability in case of server failures.