Should have called it Q3 release but whatever... Looks like they've integrated SQL Health and History tool and fixed Database Mirroring for starters.
Nothing quite like cutting it fine on the naming still being valid :-)
Get all the info and downloads from here. A quick scan reveals the following improvements;
T-SQL - Separate Data and Time Types FINALLY, YIPEEEEEEEEEEEEEE
SQL Server 2008 introduces new date and time data types. The new data types enable applications to have separate date and time types, larger year ranges for date value, larger fractional seconds precision for time value, time-zone offset aware datetime type that containing date, time and time zone offset portion, user defined option on fractional seconds precision of time related types and datetime2 and datetimeoffset provide standards conformant semantics. Along with the T-SQL support on the new types, both native (ODBC, OLEDB) and managed (SqlClient) providers also provide the full support through the client driver APIs.
T-SQL - Object Dependencies
The object dependencies improvement provides reliable discovery of dependencies between objects through newly introduced catalog view and dynamic management functions. Dependency information is always up-to-date for both schema-bound and non-schema-bound objects. Dependencies are tracked for stored procedures, tables, views, functions, triggers, user-defined types, XML schema-collections, and more.
T-SQL - ORDPATH
ORDPATH improvement provides an important new functionality to our customers who use hierarchical data. It provides a superior way of modeling hierarchies in SQL Server by introducing the HierarchyID system data type and corresponding built-in methods which are designed to make it easier to store, query and operate hierarchical data. HierarchyID is also optimized for representing trees, the most common type of hierarchical data.
T-SQL - Large User Defined Types
Large user-defined types allows users to expand the size of defined data types by eliminating the 8‑KB limit.
Reporting Services - New Server and Designer
Improvements represent the two major infrastructure changes for Reporting Services. Reporting Services enhances the processing engine and rendering extensions to enable new functionality, such as Tablix support, and scalability as well as remove the dependency on IIS. Additionally, new report designer and configuration tool are provided that improve usability and workflow for RS customers.
Management - Performance Data Collection
Collect data from various sources in SQL Server and OS to help with performance troubleshooting and server maintenance. With this improvement, organizations improve their analysis of common performance issues:
- Define what data is collected and organize the collection into collection sets
- Start/stop/manipulate collection sets programmatically (T-SQL and .NET API)
- Define where data is stored (relational database)
- View data through reports in SQL Server Management Studio.
- Provide platform to plug in more data collectors in the future.
HADR - Database Mirroring Enhancements
SQL Server 2008 builds upon the momentum of SQL Server 2005 by providing a more reliable platform with enhanced database mirroring:
- Automatic bad page repair - allows the principal and mirror machines to transparently recover from 823/824 types of data page errors by requesting a fresh copy of the corrupted page from the mirroring partner.
- Log stream compression - compression of the outgoing log stream in order to minimize the network bandwidth used by database mirroring.
- Miscellaneous performance enhancements:
- Using asynchronous log write requests on the mirror in order to shorten the log write time and thus speed-up the commit acknowledgement.
- Better utilization of the mirroring log send buffers in order to pack multiple smaller log blocks into a single network send.
- Supportability and diagnosability improvements:
- Additional performance counters to allow for more granular accounting of the time spent across the different stages of the DBM log processing.
- New DMVs and extensions of existing views in order to expose additional information about the mirroring sessions.
Brian let the cat out of the bag earlier today, then Doug confirmed it. The project I came over to VSTS to focus on is starting to get shared. As Brian discusses in his post we demo'd live code from the Aug CTP and the CTP that will follow Aug this week to our Technical Field and it was very well received. Brian's right it has been pretty challenging for the team to be working on 2 full versions of VS at the same time(VS 2008 and Rosario) but I think the team have done a great job.
We'll start talking about the details of whats in the product over the coming weeks and months but I wanted to kick off my Rosario blogging by talking a little bit about the way that we think about testing and testers. Brian refers to Manual Tester in his blog but we really think of this as short hand for testers that do manual testing (the difference will become clear in a moment).
The testing that VSTT supports today is really in 3 forms;
- Unit Testing, this is really a developer orientated form of testing and almost no "testers" use it directly but they wish all the developers in their team did!
- Web Testing, this confused me when I first joined the team as I assumed it was UI testing for web apps, its not that and I like David Williamson's definition, its about using the browser to capture the (http)data required to test an http server, when we playback(or run the tests) we do so at the http level. Testing server protocols (even http) is a pretty technical and specialised task.
- Load Testing, here we consume other tests and use them to simulate load against servers, web servers, reporting servers etc etc. We can use any test that includes code or web tests, this is somewhere that testers do/can use Unit Tests as they can form the basis of a load test. Again this is pretty technical and specialised testing.
When we decided to make a big investment in testing for VSTS going fwds (the team is more than 3x the size it was for VS 2005) one of the things we did was to segment the potential customer base into 3 broad buckets (with specialisations within, but we'll cover those in another post). Those buckets are;
- The Developer Tester, this includes both TDD and traditional Unit Testing based developers who are focussed on using dev orientated testing to increase initial and ongoing code/application quality.
- The Specialist Tester, this includes Load Testing, Security Testing, Protocol Testing, Very Deep/Complex Automation Testing etc etc, these are testers focussed on the more technical and specialised aspects of testing.
- The Generalist Tester, these testers are about making sure the functional quality of the software is right for the user community, they might be former users of the software in question they likely do lots of manual testing, scenaro testing and some form of automation(but not all organisations do automation).
We want to provide a complete testing solution for customers, and that means have products/solutions for each one of these buckets(as well as managers, leads etc). As we start to reveal the layers of Rosario we'll come back to these roles and talk about the different pieces that we are building and how they address the requirements of each.
At the World Wide Partner Conference (WWPC) this week Kevin Turner announced a launch date for the 2008 Wave of Products(Longhorn, Orcas and Katmai). There has been a lot of conjecture about products slipping and stuff like that.
To be clear this is all about a marketing launch, the RTM dates for these products are not the same as the launch date so you can draw almost no conclusions from that other than we are going to have lots of events and giveaways in Feb of next year :-)
For example when we launched VS2005, SQL2005 and Biztalk 2006 in Nov of 2005, SQL and VS had both been on MSDN for a number of weeks and Biztalk was still in Beta. When we launched Vista and Office 12 both had been available for a number of weeks before hand.
In the words of Cpl Jones "Don't panic Capt Mainwaring"
UPDATE: Doug (now in VSTS Marketing Yay!!!) says that VS will RTM before the launch, as he's in maketing he is allowed to say stuff like that, being in Prod Dev I'm not :-)
(If you don't want to read the blurb just download)
The SQL Team has had a long flirtation with building a Data Warehouse/Analytics system on top of the system information that exists in the Database Engine. The first time I remember discussing this was just after SQL 2000 shipped, we had mapped out all the SQL Agent jobs that needed to run, and what data to collect. The very first version of the product plan for Yukon had a Operations Data Warehouse feature listed but it got dropped pretty quickly as the scope of Yukon became clear.
Part of the reason it was an expensive feature in early days was the lack of consistent info in the server, we were going to have to use Trace, DBCC, System Tables, System Views and a bunch of undocumented stuff to accomplish a meaningful solution. However once DMVs started appearing in the product(that was not their original name) lots of ideas starting floating around again.
In parallel one of the small (I think there were 3 people that worked on the project) central test teams came up with the idea for H2( there was a predecessor but I don't remember its name and it was not broadly released, just select customers and internally) and built it, shipped it internally and then out to the public. There is interesting data in H2 but its mostly config data.
Through the personal perseverance of a couple of folks (mostly Dan Winn and Paul Mestamaker, then an intern, but others as well) with support from Dave Campbell we got a bunch of reports added to what was then the Summary View in Management Studio.
The Summary reports are great but they are more operational than they are analytical, lots of members of the community have filled in the gap by providing richer reports/analysis either plugged into SSMS or as standalone tools.
But as an analytics geek I have always wanted something more. Well last week the folks from the customer advisory team shipped out DMVStats which is a reporting and analytics solution focused on DMV reported perf data. I can't think of anyone better than these guys to ship such a tool as I am sure its based on real world usage. You can get it here.
Hopefully there is more to come in this space.
This sneaked out, it does not say CTP so are we to presume this is a production version?
UPDATE: Paul says here that this is the final version (for now) and they are going to be doing ongoing updates to the rules, thats pretty cool.
More Posts Next page »