Deployment
I have a DB project I created by reverse engineering our production DB. VS DBPro is looking great! I have a few questions around deployment of database objects.
1.) I notice that VS DBPro CTP creates one file for each object in my database project (in the "...\Schema Objects\" folder). The extension for each object is .sql. We have already captured all our database objects using the VS.Net 2003 database project file extensions (i.e. .TAB for table, .KCI for ref integrity constraints, .PRC for stored procedure, etc), and we have been versioning these files in VSS for a number of years now. We would like to move to using VS DBPro, but need to know whether VS DBPro will a.) allow scripting at the same granularity (i.e. sep script for tables, KCI, logins, triggers, etc) for objects in the "...\Schema Objects\" folder, and b.) whether VS DBPro will allow us to use the same file naming convention and file extensions that the VS.Net 2003 Database Project did (i.e. .TAB for table, etc). Is this something that VS BDPro will be able to do?
2.) When we deploy changes (from dev to test to production), we create a distributable package containing both a folder (with one file per object) and a .CMD file (generated using VS.Net 2003 Database Project). Deploying a .CMD file and seperate script files allows several benefits, including a.) Control: We have very granular control over what order scripts are compiled into the DB, which has a huge impact on install execution time (for example waiting until after a table schema update (create new table schema, insert existing data into new table, drop old table) is applied before applying indexes), and b.) troubleshooting: testing teams can immediately identify the script (and therefore object) that causes a failure during install. For example, if a syntax error exists in a file, the name of the file is the same as the name of the object that caused the error).
In the CTP, I notice that the install script that is generated (one big file). In VS DBPro, is there a way give me control over what is compliled when (in order) and good way to pinpoint which objects or data in the script might be causing install issues?
3.) When I deploy changes in an install package today, I am able to include objects in my install that have not changed (for example, a "work" table that needs to be dropped and rebuilt during my install). When I do a schema compare in the VS DBPro CTP, I seem to be unable to include anything but objects that have changed in the deployment ("update") script. Is this user error on my part or will the VS DBPro product allow me to include any objects I want in the my deployment in the future?
4.) I need to make sure that I can package my changes so that it cannot be changed by anyone (we don't allow anyone to decide, on go-live day, what portions of the install they'd like to deploy; the entire installation must be deployed). In addition we need to be able to compare one deployment package to another (verify differences between patches or verify that what we originally dropped is what is about to be installed into an environment). Other than instructing DBAs to deploy directly from the VS DBPro development environment (using a schema compare), What deployment mechanisms will the VS DBPro product have (will it package my scripts into a executable/.CAB file, and/or will I be able to deploy my changes using a generated .CMD file)?
Thanks in advance!
Terry
