You are currently browsing the monthly archive for February 2008.
Came across a neat tool today – I am not huge fan of Linux but I do like the ability to have multiple desktop spaces where you can divide up your work and maximise screen real estate. Well, you can now do this in Vista (or XP).
Try this out for yourself , its a little rough around the edges and it is still in beta but for those of you (like me) who like to be a little structured and organised this is a great gem of an app.
Do you need to use the excellent RMLUtilities (in particular ReadTrace) against SQL Server 2008 trace files? Well, ReadTrace does not support 2008 yet but if you need to convert your files , please try this free utility:
This is a command line tool, simply jump to the command line and use as follows:
trace.convert.exe -iC:TraceFiles (or any other folder).
Trace Convert will then convert all trace (.trc) files in this folder so that they can be used with ReadTrace. You can optionally include -r so that is converts trace files in all sub folders (recursively).
If you have a new SQL 2008 database and have used the new datatypes, this tool is not guaranteed to work as expected and you should wait until the new readtrace has been released that is compatible with SQL 2008. However, this is very handy for upgraded SQL 2005/2000 databases where you want to analyse trace output of re-playable traces.
Give it a try and see if it helps you..
One neat feature of SQL Server 2008 is the ability to see how fragmented your fulltext indexes are. By running the following you can see how many fragments you have.
SELECT * FROM sys.fulltext_index_fragments
The rule of thumb is that if you have between 30-50 fragments per table you should be thinking of reorganizing (REORGANIZE) your fulltext index.
For a clear picture, use this query:
SELECT OBJECT_NAME([table_id]) AS TableName, COUNT([fragment_id]) AS Fragments
FROM sys.fulltext_index_fragments
GROUP BY OBJECT_NAME([table_id])
HAVING COUNT([fragment_id]) >=30
Just wanted to say a big thank you to Simon Sabin and Tony Rogerson for a very enjoyable event. I really enjoyed the demo on Spatial support and Simon’s very own Spatial Viewer. We have lots of exciting times ahead with the release of SQL Server 2008.
You can download the viewer mentioned in this post from:
In case any sql dev’s wasn’t aware, TEXT and NTEXT will be deprecated. Seems like SQL 2008 could be the last version to have them, in any case I guess we should all think now about changing columns to VARCHAR/NVARCHAR(max) sooner rather than later.
This is potentially very useful if you are intending to use Change Tracking. It is possible to set a context (“hey, all these changes were done as part of this process”). This can easily be done by
DECLARE @originator_id varbinary(128);
SET @originator_id = CAST(‘SSIS’ AS varbinary(128));
WITH CHANGE_TRACKING_CONTEXT (@originator_id)
UPDATE MyTable
SET MyLocationId = 1100
WHERE MyId = 634
I say potentially because there is a drawback.. if there is another update before you bring back the changes , it’s change context is wiped.
Well, kind of. The guys behind the fantastic Camtasia Studio (techsmith) has released a fine free alternative which seems to be in beta at the moment. The product is called Jing and after play testing it for a while it seems to be very snappy. It can record video (into SWF files) or capture images. It can then send these up to Screencast (another Techsmith product/service), FTP to a location of your choice, post up to Flickr or save to file. It is very basic but because of this it is extremely easy to use.
For those of you who can’t afford camtasia , this is a very nice introduction and will probably meet the majority of people’s needs.
Jing can be downloaded from:
I have been doing some benchmark/tracing for Katmai and came across the excellent RML Utilities. However I wanted to blog about a bug in this which has meant that the ostress tool couldn’t be used to replay the trace.
ReadTrace allows you to read a trace file and output/split this into RML files (one per spid). You can then do some fancy stuff with ostress to replay these files. However, readtrace appears to have a bug where it incorrectly parses your trace file if you have used named parameters in your rpc calls.
This is an example call from a trace file
exec up_Table_Search @tableid=1, @namedpar1 = N‘test’, @namedpar2 = 2, @namedpar3 = N‘test2′
When this is parsed by readtrace, it will produce a RML file that calls the proc like this:
exec up_Table_Search 1, ‘test’,2,‘test2′
Why is this a problem? Well, what if the order of parameters isn’t the same in the proc..
ReadTrace does not seem to cater for named parameters and products an RML file which can potentially either cause data corruption or fail to run at all if the datatypes are incompatible.
Please , someone let me know if there is a command line switch to fix this. There doesn’t seem to be.
