In June 2006, Microsoft published a SQL Server technical paper on Physical Database Storage Design. This paper was updated in February 2007. The paper is generally well written, and the recommendations are reasonable. However, the following two specific recommendations caught my attention: For small servers with less than three disks ...

There has been much discussion on the usefulness of disk queue length as an indicator of a disk I/O bottleneck. Bob Dorr, for instance, addressed this issue directly in his excellent blog, SQL Server Urban Legends Discussed. But the issue is not settled for many since we still frequently see people using disk queue length as the key measure in ...

The best documentation on the I/O behavior of SQL Server checkpoints is found in SQL Server 2000 I/O Basics by Bob Dorr. In particular, you should read the following carefully: SQL Server uses the following steps to set up another page for flushing and repeats for up to 16 total pages inclusive of the first page. Do a hash lookup for the ...

Too many DBAs tend to view a drive presented from a Storage Area Network (SAN) as something of a monolithic nature. They look at the drive as if it had some intrinsic performance characteristics. This view doesn't help one appreciate the true performance characteristics of such a drive. A more constructive view is to look at the drive as an I/O ...

With the insert script and the test configurations in my previous posts, the best data load throughput was 24GB in ~7 minutes when the checkpoint (and/or transaction commit) batch size was set to 100,000 ~ 1,000,000. That was the best result when I was trying to get the most out of the insert script. But if we forget about tinkering the ...

In my previous posts (1, 2, 3), I focused on the performance behavior of setting the checkpoints and transaction commit sizes to once every 16 inserts and once every 100,000 inserts. A question remains: what is the most optimal size? In other words, in the following script, what value should we give to variable @batch_size so that the script ...

In my two previous posts on the performance impact of frequent manual checkpoints and the I/O behavior of frequent manual checkpoints, I demonstrated that frequently issuing manual checkpoints can be bad for performance and why it's bad from the storage perspective. If you were led to believe that manual checkpoints were always bad, that wasn't ...

In my previous blog post on the performance impact of frequent manual checkpoints, I highlighted the performance peril of going overboard with manual checkpoints, and I suggested that a major contributing factor was the failure of frequent manual checkpoints to take advantage of the throughput potential of the underlying storage. But I didn't ...

In SQL Server related storage literature, it is almost univerally recommended that disk partitions be aligned either on the 32K or the 64K boundary. On this subject, I posted some test results a while back, regarding the performance impact of disk partition misalignment: ...

SQL Server 2005 supports placing read-only filegroups or read-only databases on NTFS compression. In other words, you can compress the database files in a read-only filegroup or a read-only database. This can be a very useful feature if saving disk storage is of high priority. But what are the performance implications of using this SQL Server ...