April 25, 2007 at 6:55 pm
Good afternoon everyone,
I am fairly new to this stuff and am trying to get a solid understanding of how file groups work, how best to use them and what is really going on in the background.
As I understand it, when you create objects in the DB, those objects will be created in the default filegroup (PRIMARY), unless you have created additional filegroups and explicitly designated the objects be created in a different filegroup.
I've read that it is a good practice to always create a 2nd data file *.ndf, associating that file with a new filegroup (making this new filegroup the default for all newly created bojects) such that the only information in the *.mdf file in the PRIMARY filegroup just has a pointer to the 2nd file?
Are there white papers on filegroups that I can read?
Any information on this subject would be greatly appreciated.
April 25, 2007 at 8:18 pm
There is a good first place to look at MSDN. I haven't figured out how to make the link live.... ;o(
http://msdn2.microsoft.com/en-us/library/ms179316.aspx
Definitely worth a search for "Physical Database Architecture" in the BOL as well.
It's good practice to have your OS and SQL Server log files on a different physical hard drive than your data files (and perhaps to have data files on different disks than each other) if you have a write-intensive OLTP database). We can't all afford dual-bus architecture, but it's good stuff if you can get it. ;o)
If you aren't experiencing an abnormal disk load, separate filegroups might not be necessary. As far as your paragraph beginning with "I've read that..." I need to let someone else handle that, because not heard of that and don't know what impact it would have, good or bad.
:o)
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply