question. i have a table that is around 800 million records. i was using partitioning w/ an incrementing identity column to split it across 8 file groups (8 different disks) but far-passed the qty for the last disk, so now all new data is on the last disk. is there any way to design the table/db in a manner so that once all of the disks in the partition are full, it starts back at the beginning of the chain? i basically would like it to evenly distribute the data across all 8 disks and auto grow. probably a really NUB question, and a simple answer, but not sure what to do here.
This is not my area of expertise, but shouldn't you have used one filegroup spread across the disks? That's how we do it here. Just not sure how that fits into partitioning.
I'd have one filegroup which would have 8 different files, one file on each drive/mount point. SQL Server uses a proportional fill algorithm to spread the data across them. They don't need to evenly sized like they do for tempdb, but I do keep them at the same size. SQL just tries to the keep the full percentage the same across them, regardless of size.
can you give me a sample db create script for one file group, two files with two separate mount points? i'm playing with the wizard and i'm a little lost, the script would probably spell it out for me.
file groups / partitioning questions
Posted - 06/07/2012 : 16:52:27
