December 16, 2008 at 7:53 pm
This company has put all .mdf files, .ldf files for all databases on all instances on netapp LUNS.
I don't understand the technology very well so I have yet to determine if this solution is ok or not. So here is what I know and feel free to fill in the blanks:
All disks are shared among LUN's in an aggregate.
This means that even if the .mdf files, tempdb and the .ldf files are on different LUNS, they share the same disks. If this is true, then in order to truely dedicate a certain amount of disks for a database file, the LUN has to reside on a different aggregate. Also, if this is true, I believe that that other apps such as exchange, shares this same aggregate (very scary). Can anyone verify this information? I've searched google for Netapp topics but still find myself with questions that the sales people can't seem to answer.
Next -
If this Netapp uses RAID DP which is RAID 4 with Dual Parity, then that means that I should use the following formula for to calculate IOPS:
(reads + (writes * 4))/Disks - 2
According to different forums and documents on Netapp, since RAID DP uses WAFL technology to somehow buffer writes and write then later in logical order, there really is no write penalty thus we get the same I/O as RAID 0? I don't think I'm understanding this right (thus I put this topic in the newby board).
Please help?
January 22, 2009 at 11:35 am
I have the same problem and questions. I inherited the same type of environment. I hope someone answers this thread.
Viewing 2 posts - 1 through 1 (of 1 total)
You must be logged in to reply to this topic. Login to reply