Posts

I was setting up mirroring at a client site today and started as I usually do by backing up the primary database and restoring it at the mirror. It would not start mirroring and complained that I didn’t have the latest transaction log backup: the mirror database …., has insufficient transaction log data to preserve the log backup chain of the principal database.

Somewhere along the way, the need to have a transaction log backup has crept in. So, to start mirroring, I just now backup the primary database to a file, backup the transaction log to the same file and then restore both on the mirror server. Then it starts fine.

2008-08-18

Recently, I was working at a client’s site where SQL Server 2000 was still in use. They wanted to upgrade to SQL Server 2005. The only reason they were looking at the Enterprise Edition was for data-driven subscriptions in Reporting Services. The CEO certainly wasn’t interested in paying the sort of dollars required for that, just to be able to automatically email out a set of monthly reports to about 30 or so folk within the company.

2008-08-13

While I was prepping my full-text search session for TechEd Australia today, I decided to modify the thesaurus. I found the discussion in books online a bit confusing regarding the locations of the files involved.

What threw me was it said the default location of the thesaurus files was:

_SQL_Server_install_path_\Microsoft SQL Server\MSSQL10.MSSQLSERVER\MSSQL\FTDATA\

I went there and couldn’t find anything. I worked out that for ENU language, my file should be called tsENU.xml. I searched for that and found it in:

2008-08-13

I’ve been doing another performance tuning job today and it highlighted to me once again that problem that a lack of indexes on foreign key columns can bring.

By adding indexes on the foreign keys on three tables, we saw a reduction of 87% in total I/O load on the server. There are other aspects of the system that I’m now working on but it really struck me that having SQL Server do this by default would avoid a lot of apparent performance problems. It would have to be one of the most common indexing issues that I see in my work.

2008-07-28

I’ve also been looking around for a new phone. It became much more urgent last week after I dropped my iMate.

After trying lots of phones, I ended up opting for an iPhone 3G. And after using it for a few days, couldn’t be happier with it. Well, not quite true, if it worked as a NextG modem as well, it would be even better as the data plans are quite costly here in Australia.

2008-07-22

Hi Folks,

SQL Down Under show 38 with guest SQL Server MVP Andy Leonard discussing database testing is now available for download from https://sqldownunder.com .

The show also includes a competition.

Enjoy !

2008-07-19

I caught up with James Luetkehoelter at the PASS Summit in Germany a few months back. He sent me a copy of his new book from APress: Pro SQL Server Disaster Recovery .

I managed to finish reading it while heading back from CodeCampSA in Adelaide today (which was a good solid event again - excellent work Peter Griffiths!). I quite enjoyed the book and I like James’ writing style. It’s quite conversational and I could hear him talking to me as I read it.

2008-07-13

There are a number of key books that I’ve missed reading over the years, in areas that interest me. Recently, I’ve been fixing that. One that is always discussed is The Microsoft Data Warehouse Toolkit by Joy Mundy and Warren Thornthwaite from the Kimball Group.

I would have to say I enjoyed reading it. It is a large book at over 700 pages and a couple of inches thick so it took a while to get through.

2008-07-09

Well I was called out by Tibor Karaszi’s meme, so here goes:

How old were you when you first started programming?

I’d say I was about 19 when I started. I remember in 1976 that I was at University of Queensland. I was doing an honours degree in physics and maths and didn’t have the slightest interest in those computing people that spent their lunch hours looking at great piles of 15x11 listings. By the next year, I was one of them.

2008-07-08

I know this post has the potential for religious-level debate but it’s time to make it anyway.

The more I’ve been working with Analysis Services lately, the more it grates on me that the BI community still seem to be the last ones hanging onto table prefixes. They’re not doing “tblSomeTable” but they are using “dim”, “fact”, etc.

Hasn’t the time for this long gone now?

Most of the argument seems to be about finding tables in a list of tables. You could do that via schemas if you really wanted to. But as Adam Machanic pointed out recently, from 2005 onwards many-to-many dimensions blur these lines anyway.

2008-06-16