Sie sind auf Seite 1von 3

ProProfs.

com [CONFIGURING LOG AND DATA FILES IN SQL SERVER]

Configuring Log and Data Files


Any SQL Server 2005 database will have two file types: data files and log files. A data file or a log file can
only belong to a single database.

 Data Files hold data and objects (tables, indexes, etc).


 Log Files hold the transaction log used for tracking/recovering the database’s transactions.

Data Files

Data Files can be further broken down into two separate types:

1. Primary Data Files are always included in a


database; they contain startup information for
the database, and contain pointers to the other
files in the database. Can also hold objects/user
data. The “suggested” extension for primary data
files is “.mdf”. Each database can only have one
primary data file.

2. Secondary Data Files are optional and user-defined. These contain objects and user data. When
administering secondary data files, you can place secondary files on different physical disk drives
for a boost in performance. The “suggested” extension for secondary data files is “.ndf”. Each
database can have zero or more secondary data files.

Best Practice: To reduce disk access contention, you should place all data and objects in secondary files
and keep just the database catalog in the primary file.

Log Files

You must have at least a single transaction log for every database, but you are allowed multiple log files
per database to provide for faster recovery. The “suggested” extension for log files is “.ldf”.

Filegroups

A filegroup is simply a logical structure that allows the grouping of data files for management of multiple
files as a single logical entity. For a performance boost, you can divide database objects across several
filegroups, and place the filegroups on different disk subsystems.
ProProfs.com [CONFIGURING LOG AND DATA FILES IN SQL SERVER]

 Primary Filegroup: contains the primary data file and any secondary data files that aren’t
explicitly stored in another filegroup.
 User-defined Filegroup: Created to group secondary files, assign database objects to groups.

Every database has a default filegroup


(named PRIMARY), and unless a filegroup is
specified, the object is assigned to that
filegroup. To alter the database’s default filegroup, issue the T-SQL statement:

ALTER DATABASE [target database] MODIFY FILEGROUP [filegroup name] DEFAULT

Configuring Data Files and Log Files

PRIMARY: Specifies a primary


filegroup

NAME: The logical name of the


filegroup

FILENAME: The path to the filegroup

SIZE: The size of the file.

MAXSIZE: The maximum size you will


allow the filegroup to grow

FILEGROWTH: Specifies the


automatic growth interval for the file.

LOG ON: Specifies information for the


logfile.
ProProfs.com [CONFIGURING LOG AND DATA FILES IN SQL SERVER]

To edit a filegroup:

To edit a filegroup, specify the database you


want to alter, issue the “ADD FILE”
command, specify parameters like you would
normally, and then specify to which filegroup
you want to add the datafile to.

Configuring Database Files with RAID Systems

Using RAID subsystems provides for improved performance, fault tolerance, and more storage capacity.

- RAID 0: Disk striping, best performance for read/write operations; no fault tolerance.
- RAID 1: Disk mirroring, provides fault tolerance; improves read performance, but can degrade
write performance.
- RAID 5: Disk striping with distributed parity. Improved performance over RAID 1, but more
expensive.
- RAID 10 (RAID 1+0): A mirror of a striped set (combines both speed and fault tolerance)

For a comparison on RAID levels, visit: http://msdn2.microsoft.com/en-us/library/ms178048.aspx

Overview of Best Practices

- Do not put data files on the same drive as the OS files.


- Place transaction log files on a different drive than the data files.
- Place tempdb database on a separate drive (RAID 10 or RAID 5 system).

Das könnte Ihnen auch gefallen