The default setting in SQL Server 2005 and later is to keep a trace running all the time that captures basic information about the instance. This is called the Default Trace. The created trace files can be viewed in Performance Monitor or by using T-SQL and they can be very useful in some troubleshooting scenarios.
When you set up SQL Server alerts on your SQL Server instance, make sure to run some tests to verify that they are actually firing when errors occur.
Database Snapshots is an Enterprise feature introduced in SQL Server 2005, and represent a database and its data at a fixed point in time. Among other things, they can be useful for reporting scenarios and as a restore point (for instance after a major update).
On a machine running SQL Server there are a lot of performance counters that can be used to create baselines or to investigate performance problems. Those specific to SQL Server can be sampled regularly from the system table sys.dm_os_performance_counters. But those belonging to the OS must be gathered in some other way. Since I’m a SQL Server DBA, of course I want the counters saved to a table in SQL Server 🙂
To get all inserted or updated members from a SQL Server MDS database, subscription views are used. Deleted members can’t be found in the views, so another technique must be used.
In an earlier blog post I demonstrated how to map between hierarchical (XML) and relational (SQL) data using Altova MapForce. I also mentioned that MapForce can generate code (C#, C++, XSLT for instance) to execute the mapping in an environment outside MapForce. This being a blog mostly about SQL Server, one obvious question is how to run the generated mapping from SSIS.
In an earlier blog post, I demonstrated a way to insert a hierarchical XML document into a relational table structure by using XQuery. That solution works well in smaller solutions, but it’s a “code only” solution that can become complex and hard to maintain in larger projects.
An alternative solution would be to use the graphical mapping tool Mapforce from Altova.
The min and max aggregates in SQL Server are normally used to work over rows, getting the max or min value from a set of rows. But what if you want to get the max or min value from a set of columns instead?
