How to control the excessive use of ram by SQL Server?

What is the usual methodology for using Git when developing in SQL Server (stored procedures, triggers, etc.)?

• Thanks for your answer. The question refers to using git (the versioning system) to version control and collaborate on SQL development. Since SQL files don't have an effect until they are compiled as DB objects, it's not straightforward to me as to how to employ git. What is the customary methodology to use git? How are variations in table structure managed, as well?

• Answer:

  You need to think beyond version controlling individual files to how you are going to version control and manage your database schema in production, development or staging environments.  You could consider your stored procedures as schema elements, or related to the schema. So the issues are around managing change control of your schema  elements as you need to update them for new software releases, data changes and bug fixes. Consider your database schema as having a version level. So you have deployed schema v1 and you are developing v1.1 which is going to have some updates to tables, views, stored procedures etc.  Schema v1.1 is going to be some sort of 'change set ' applied on top of v1. You can manage this with some tools like Liquibase (http://www.liquibase.org) or Flyway (http://flywaydb.org). The Liquibase tool allows you to manage the change sets as XML configuration for example, referencing stored procedure definitions as SQL files. So these would be the files you would be committing to Git or any other source control system in a specific directory structure you need to define for the database release change sets. These tools typically have a way of tracking which changes have been applied to the database and only updating what is required, using tables they create in the database to track this. They can also output the updates as plain SQL for a DBA to review prior to deployment. Once you have this in place you can look at running tests in a CI tool such as Jenkins to automatically deploy the schema updates into a Dev or Staging database environment and run tests to smoke out any issues arising during development. We used Liqiuibase on a project I worked on and it made deployment of schema updates much more predictable replacing an ad hoc manual process that was prone to error. We also ran unit tests against all our stored procedures using Groovy, Gradle and JUnit.