Showing posts with label span. Show all posts
Showing posts with label span. Show all posts

Monday, March 19, 2012

filegroups

Hi,
Due to the size of our tables we're considering the use of filegroups that
span multiple disk-arrays.
What's the best aproach:
- create a filegroup with files on multiple disks an let SQL Server figure
out/spread the table data across the disks automaticly.
or
- Analyse and assign the tables to groups/files/disks ourselves
tanx,
Derk JanDerk,
Create your user-defined filegroup(s). Change the default filegroup status
from PRIMARY to one of your user-defined filegroups (ALTER DATABASE). Use
the ON FILEGROUP clause of the CREATE INDEX and CREATE TABLE statements to
control the placement of your tables and indexes.
HTH
Jerry
"Derk Jan" <DerkJan@.discussions.microsoft.com> wrote in message
news:9A2C5762-98F9-470C-95CE-50F1DCDA05DC@.microsoft.com...
> Hi,
> Due to the size of our tables we're considering the use of filegroups that
> span multiple disk-arrays.
> What's the best aproach:
> - create a filegroup with files on multiple disks an let SQL Server figure
> out/spread the table data across the disks automaticly.
> or
> - Analyse and assign the tables to groups/files/disks ourselves
> tanx,
> Derk Jan
>|||Why does the size of the table warrant spanning multiple arrays? Are you
scanning the entire table each time you access it? How large are you
talking about? The size or number of rows in a table or database means much
less than how you use and access the data. Before you go moving stuff
around you should determine where the bottleneck really is and why. That
will play a major role in deciding how to split up data if required.
--
Andrew J. Kelly SQL MVP
"Derk Jan" <DerkJan@.discussions.microsoft.com> wrote in message
news:9A2C5762-98F9-470C-95CE-50F1DCDA05DC@.microsoft.com...
> Hi,
> Due to the size of our tables we're considering the use of filegroups that
> span multiple disk-arrays.
> What's the best aproach:
> - create a filegroup with files on multiple disks an let SQL Server figure
> out/spread the table data across the disks automaticly.
> or
> - Analyse and assign the tables to groups/files/disks ourselves
> tanx,
> Derk Jan
>