February 6, 2009 at 11:54 am
I would swear that in some book/blog/seminar I had received the message that adding an additional data file after the fact to a very large database was asking for a performance hit of some kind (maybe fragmentation?) but I can not find anything to answer the question so here is my situation:
We have a somewhat large (~900 GB) database that is currently spread across two 300 GB data file and a 300 GB index data file...we are at the point where we need to grow and we are discussing whether there will be a performance hit one way or the other to adding an additional 300 GB data file, probably on a new LUN, versus expanding the current LUN's to allow for a similar growth of the existing data files.
Any insights? Thanks!
AndyG
February 6, 2009 at 5:03 pm
AndyG (2/6/2009)
adding an additional 300 GB data file, probably on a new LUN, versus expanding the current LUN's to allow for a similar growth of the existing data files.
I would think adding an additional data file to a new LUN would provide a small performance gain if you have a multiple processor machine. This allows another thread to access the new datafile and adds disk heads. If you add another datafile to an existing LUN you won't see the performance gain because you still have the same number of disk heads under the LUN. MS recommends having .25 to 1 datafile per CPU.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply