There are many cases where this can be useful - for starters, for reasons of accessibility, it is always better to store only system data in your main data file (with Sql2k5 and higher, if the main data file is available, the database can be connected to the network, allowing you to restore / restore / etc. non-system data, having as many online features as possible). Some other uses of secondary files:
- Splitting data across multiple LUNs
- Allow partial / backup / restore filegroups
- Segmenting different types of read / write accesses on different LUNs (e.g. sequential or random)
source
share