Is there a design philosophy that permeates the new and improved management tools?
Our guiding principles for SQL Server 2005 tools are "no secrets" and integration. "No secrets" means that everything we do in the UI should be exposed to users and independent software vendors (ISVs) through an API and that DBAs should be able to work out what we've done and be able to reproduce it more easily. For integration, we built SQL Server Management Studio to aggregate the management of three different components—SQL Server, Analysis Services, and Reporting Services—and several other components. In addition, SQL Computer Manager, Profiler, and Replay work with more than just SQL Server.
Can SQL Server DBAs and developers keep Enterprise Manager and Query Analyzer on their desktops and use them as they're getting used to the new SQL Server 2005 tools? And what is Microsoft doing in the toolset for backward compatibility?
We want to make this an easy transition for customers. So if you install SQL Server 2005 alongside SQL Server 2000, existing versions of Enterprise Manager and Query Analyzer will continue to service SQL Server 2000 and 7.0. However, if you upgrade to SQL Server 2005, we'll replace the tools. Management Studio also works against non-SQL Server 2005 servers, as does Profiler and other tools.
Will SQL Server 2005 or a future release provide built-in tools for data loading and stress testing?
The tools I use to perform these tasks are DTS or SQLCMD for data loading and Replay for stress testing (this is the same tool we use internally). The new capabilities in SQLCMD make it a powerful tool for batch work, including data loading. And in Replay, you can rerun traces from custom managed-code applications because we've exposed the Replay engine through a series of APIs that let you perform the same actions without having to use the Replay GUI.
Do SQL Server 2005's management tools let you isolate resource-utilization problems, such as excessive I/O, CPU, and memory utilization, and who's causing these problems? Currently, customers have to use undocumented features to isolate such problems—for example, using DBCC SQLPERF with WAITSTATS, IOSTATS, and UMSSTATS options. What is Microsoft doing to help customers tune and analyze SQL Server more effectively?
No matter who or what is causing the resource-utilization problems, the first step to solving them is making sure that the data is available. With the next release of SQL Server, we've focused on providing more information within the servers. For the relational server, you can see this data through a new technology called Dynamic Management Views. These views provide a snapshot of internal structures and states so that you can query, archive, and compare such information. We'll expose this information through a series of reports, but we're also working with tools vendors to help them better expose and leverage this information. Exposing this information will make the server far more transparent and will help customers diagnose many of the issues you list.
Does Microsoft plan on adding a SQL Server "health" application that can evaluate and suggest tuning enhancements and configuration changes?
Today, customers can use the SQL Server 2000 Best Practices Analyzer (BPA) to scan their systems and ensure they're up-to-date. BPA is in beta testing now, but already, we've received positive feedback about its ability to help customers make their systems more productive. In addition, we work closely with the Microsoft Operations Manager (MOM) team and, last year, released a major update to the SQL Server Management Pack that's part of MOM. MOM 2005 will feature significant enhancements, including the explicit changes we engineered to support SQL Server. One example is that you'll be able to run diagnostic queries from inside the MOM Operations Console.
A key change that will make the next release of MOM more effective for SQL Server professionals is that it will include fewer SQL rules by default. This might seem like a weird enhancement, but bear with me for a moment. One challenge in an operations environment is prioritizing problems. Because of this, when a warning goes off, it has to be real. In MOM-speak, this means making sure that when we raise an alert, it's valid and critical to the environment. We've also worked on adding SQL Server-specific tasks to the environment.
Prev. page
1
[2]
3
4
next page 
