I am looking at putting together a SQL server, and I am getting some contradictory info from dell.
It has been best practice for a long time to separate OS/SQL Data/SQL Logs/TempDB on separate arrays, but I am now hearing that OBR10 (One Big RAID 10) is the way to go. If it is not the way to go, it will at least be suitable for our environment, per dell.
I am interested in hearing an opinion confirming, or proposing some reasoning why this would not always be the case.
If the current IO load fits under the maximum potential IO of the new server is that all I would need to be concerned about, or are there more complicated reasons why I would not want to use OBR? I understood the reasoning more for competing IO characteristics, sequential vs random, and keeping them separate. Is it all IO, and there for I only need to account for current IO, and make sure the number of spindles, and speed of the disks puts my IO potential well above that?