May 27, 2015 at 10:27 am
Hi,
I plan to implement the disk spanning feature in Windows Server 2012 in order to create a single 12 TB volume (using 2 x 6TB disks). The new volume will hold a 10TB database while using SQL Server 2012.
The question is: Are there any known performance or stability issues when using SQL Server 2012 with the Widows Server disk spanning feature?
Regards,
-r
May 27, 2015 at 12:10 pm
raf.figueroa (5/27/2015)
Hi,I plan to implement the disk spanning feature in Windows Server 2012 in order to create a single 12 TB volume (using 2 x 6TB disks). The new volume will hold a 10TB database while using SQL Server 2012.
The question is: Are there any known performance or stability issues when using SQL Server 2012 with the Widows Server disk spanning feature?
Regards,
-r
The answer to your question is irrelevant. You have NO HOPE of decent performance running a 10TB database on 6 disks!!!!!!! Do the math: 100-150MB/sec on sequential IO (WAY less if random IO) per disk. So how long would it take to scan your largest object??
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
May 27, 2015 at 1:16 pm
Kevin,
I am not sure where did the 6 disk come from? in my original posting the scenario is to use 2 (two) 6TB disk, when spanned the two disk's should be seen as a single 12TB volume, where my 10TB DB will reside, does that clarify things a bit?
May 27, 2015 at 1:37 pm
In that case, it'll be 6 times worst than hopeless.
When it comes to IO throughput, the number of spindles is critical. You have 2. My laptop has that many in a 1TB spanned drive, you don't want to run a production DB on it as it is terrible with IO. 🙂
The other problem here is redundancy. When one of those drives fail, good bye database. That's why important databases are usually put onto RAID arrays, redundant arrays of disks so that when one drive fails the system carries on running fine.
Hell, both my home desktop (gaming machine mostly) and server (test server + DC) have the drives in a RAID array.
Honestly, you probably want to rethink things. Look at larger numbers of smaller drives and look at RAID 5 or 10.
If you insist on going ahead with this, I recommend you work out what portion of that 10TB database is active and used, and put about 1.5*(active portion of the 10TB database) of memory into the server. You might get usable performance that way.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
May 27, 2015 at 3:40 pm
raf.figueroa (5/27/2015)
Kevin,I am not sure where did the 6 disk come from? in my original posting the scenario is to use 2 (two) 6TB disk, when spanned the two disk's should be seen as a single 12TB volume, where my 10TB DB will reside, does that clarify things a bit?
Yep, and as Gail said it makes things even worse than I dyslexically saw them as! :w00t:
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
May 28, 2015 at 4:48 am
Also don't forget that you don't just bottleneck on the disks. You also bottleneck on the disk controllers and the connections to the disks. I/O is a giant rats nest and with the scale you're talking, you're going to have to get serious about it. I'd suggest getting a hardware expert to help out on this one.
"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt
Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply