3

I have a SQL Server setup that a previous IT person set up with a 2TB data partition and a 1TB log partition. The OS partition is 244GB and SQL is installed on a separate 1TB partition.

We have an additional 8TB of storage that I would like the new IT staff to bring on line. He wants to create 4 new 2TB data partition. I see this as confusing. Can't we just backup the current data partition, blow it away, and create a new 10TB data partition

I'm responsible for administering the data on the server but am not allowed to do the setup myself. This is a GIS server running ArcGIS Server with around 60 geodatabases ranging from 20BG to a couple that may grow to over a TB.

So, 5-2TB data partitions or 1-10TB partition.

Thanks for the advice.

Reese Hirth
  • 31
  • 1
  • Why don't you ask "him" why he wants to do it this way? It may be a limitation of the storage infrastructure. It may also be a limitation of the platform. For instance, if your server is a VMware or Hyper-V VM then I believe the limit for a VMDK or VHD is 2TB. – joeqwerty May 30 '12 at 14:37

2 Answers2

1

I think that whenever you add a considerable amount of storage (and I think adding 300% is sizeable) it's time to re-evaluate how storage is used as a whole.

There are partitions and there are drives, so the first thing you want to do is verify that the actual drives used with each partition are separate. (note that some storage arrays do not allow this- or make it difficult to do- like 3par and HP EVA arrays) If they are not separate you still need to know where you stand. Databases should not be 1 single file. Databases should have a number of files equal to the number of cores available to SQL server. These files can be spread across multiple partitions. Additionally TempDB should be located on your fastest drives (or at a minimum I'd dedicate a partition to tempDB.

These are some general reccomendations - for more details see Physical Database Storage Design

Jim B
  • 24,081
  • 4
  • 36
  • 60
0

The idea is to keep the SQL data and log files on separate discs and spindles to follow SQL best practices.

The idea is that this will give you a much higher IO as it stops contention for the disc from the data or log.

http://technet.microsoft.com/en-us/library/bb402876.aspx

I would recommend at least having two different partitions set on different sets of discs to make sure the data and log are definitely on different discs and spindles.

Also, for the partition used to host SQL data I would definitely recommend creating your disk arrays in RAID 10 for maximum disk I/O and redundancy. For the partition created for logs, you should only need RAID 5.

boburob
  • 1,174
  • 8
  • 23