Much like Wade Wegner I think that SQL Azure is the jewel in the cloud for Microsoft. None of the other vendors have anything like it. While it can be a bit sticker-shock-ish given that nominally a gig in SQL Azure is 65 times the price of a gig in Azure storage, once you actually run some real world scenario models it turns out to be really well priced. Think of a SQL Azure instance not as 1GB (or 10GB… but I see no reason to use 10GB partitions) of storage but rather as the cheapest damn fully backed up and HA relational database solution you’ll find anywhere- and yes FOSSers I include your stuff in that calculation; no greasy haired, under washed and over WoWed engineers needed here.

Anyway… I digress.

A major PITA in using SQL Azure is that the tooling is tantalizingly close to be OK… but in many ways it just doesn’t work. Dumping a SQL script and then re-creating the DB in SQL Azure is a painful exercise in find and replace- check out the hands on lab on Migrating a DB to the cloud in the Azure training kit for the gory details…

Wade has a blog post up about a freebie tool written by George Huey that automates this process for you. Essential for your Azure kit bag. It will parse out all the unsupported stuff.

I ran it up and gave it a nice brutal challenge…. the AdventureWorksLT script that’s used in the aforementioned lab. This includes both schema and data and is a decent effort to parse. The tool churned for a good 3 or 4 minutes…. but I got a script out! The original script includes some real curve-balls like XML Indexes and some tables with data to populate but no clustered index (SQL Azure needs a clustered index before you can insert into a table). I don’t expect it’ll get everything right., but, let’s take a look at how it does against my hand crafted script…..

DOH!

It doesn’t support cut and paste or saving of the script yet so I’ll need to go and change the source first… More reporting back from me later. This looks really promising and I’m confident it’s going to solve 90% of the pain points I’ve been hitting trying to move complex (hell even simple) databases to the cloud.