How to release a hotfix with pull-request inside VSTS in 3 steps

We all know the situation: the customer finds a critical bug in the latest release and he wants us to release a new version of our application with a fix. How do we handle this situation without breaking our team policies? How to release a specific fix to avoid regression problems?

First we need to fix that bug with the classical approach of feature-branches. We create a branch, fix the bug, create a pull-request and the team approves. These activities are set at the highest priority because a customer is in trouble and we must help.

Now we are in a situation where we have a specific commit that solves a specific problem with only the necessary lines of code modified.

commit-with-fix.png

Our target is to ship that specific commit that fixes that specific bug with a standard pull-request, without all the other work done in the development branch since the latest release. So we write down (in our clipboard for example) the id of the commit.

shaid

What can we do now to release?

1. New branch

We create a new branch.

git checkout -b my-hotifx

Now we fetch the latest updates from the remote repo.

git fetch origin

Then we set our branch my-hotfix to point to the latest commit of the remote release branch.

git reset --hard origin/release

branch-release.png

2. Cherry-pick

Now we cherry-pick the specific commit we want to apply to the release branch without commiting (–no-commit option). We choose the no-commit option to carefully inspect what is going on in our files. It’s here where we use the commit id we saved early.

git cherry-pick <commit-hash> --no-commit

We verify that everything is fine (where fine depends on your specific project). Now we can commit and push to VSTS.

git add .
git commit -m "Hotfix"
git push origin mybranch:mybranch

final-situation-1

3. Open pull-request

Our branch is now on the remote repo and we can open the pull-request with VSTS.

pull-request.png

From here the team can approve the PR and fire up our automated release process that activates our automated test and if everything is fine we deploy safely.

TL; DR

With this blog post we explored the critical situation to release a hotfix without breaking the rules or taking shortcuts to avoid our release pipeline. As engineers we must maintain a cold approach even in hot situation and rely on our best practices. Human errors are always possible in particular when we’re stressed and a customer is making our phones hot.

Advertisements

ALM DOs and DON’Ts – Unit test

A unit test is a runnable piece of code that verifies that another piece of code (called production code) does what it is supposed to do.

A unit test has many characteristics and one of the most important is that a single unit test must verify one and only one thing.
If a unit test specifies more than one behavior things became harder to mantain. One a tests fail it has to be easy to understand what is going wrong and which section of our production code is behaving badly.

unit-test-one-thing.png

As a best practice we make each test independent to all the others: any given behaviour should be specified in one and only one test. Otherwise if you later change that behaviour, you’ll have to change multiple tests.

 

ALM DOs & DON’Ts – Database source control

Every application needs to store its data. A (relational) database is the most common choice in many situations. Like every other component of our software project, a database schema must be managed with a version control system. We will never work without source control for our source code files and we must use the same mindset when dealing with a db.

database-source-control.png

The database is a critical part of our application. If we deploy version 2.0 of our application against version 1.0 of our database, what do we get? A broken application. And that’s why our database should always be under source control right next to our application code.

ALM DOs and DON’Ts – CI builds status

We all know the practice of continuous integration.

One of the common pitfalls about CI is that the build status is not monitored and not treated as one of the top priorities for the team.

build-ci-status

A healty/green status of our CI process means that our code is in a good shape for what our automated tests can tell. Fixing the build status ASAP is easier than leave it red and fix later because the recent changes of the codebase are vivid in the team members’ memory.

Quella sottile linea invisibile

Camminando giusto giusto sul confine tra Italia e Austria nel percorso Water Trail.

Un passo sei di qua e un passo sei di là. Così semplice, quasi banale. Eppure la differenza è enorme. Sarà una fissa mia ma queste cose mi affascinano. E mi chiedo: perché proprio lì la linea di confine e non un metro prima o dopo? Come viene deciso?

Questi confini geografici mi scatenano poi dei pensieri piu generici o astratti. Quanti confini attraversiamo nella nostra vita? E quanti ammettono un tornare indietro?

I nostri confini sono anche ciò che ci caratterizza: i nostri limiti, i nostri princìpi, la nostra confort zone. Cosa facciamo quotidianamente e in modo tangibile per espandere noi stessi, le nostre abilità e le nostre conoscenze?

Detto questo mi fermo qua, potrei cominciare a farneticare e sconfinare in un territorio filosofeggiante che non mi appartiene.

A voi affascinano i limiti, i confini, le situazioni limite?

Mastering database migrations scripts with VSTS

During the lifecycle of our applications it’s very likely that to support new functionalities we need to modify the structure of our database. In this particular post I mean a relational database.

When we deal with a database in a production environment we can’t simply drop an existing database and create a new one with the latest version of the schema.

The database is unlike application code in as much as it contains state, in terms of table-based data,  that needs to be preserved after an upgrade.

drop-create

This is why when dealing with database changes we need to talk about migration scripts.

Migration Scripts

A migration script is a set of SQL commands that changes the database schema or updates data to evolve the database from the version n-1 to the version whitout loss of data.

Migration scripts are typically collected in a dedicated folder in our repository, named sequentially.

The migration scripts must be executed in the specific order to be able to build the n-th version of our database.

The first migration script

The first script in our collection is responsible to create the first version of our db starting from an empty database.

The following is an example of a simple first migration script that we can name 00000001.sql. The example is for a simple database with one schema named Ordering and one table called Customers in the Ordering schema.

CREATE SCHEMA Ordering
GO
CREATE TABLE Ordering.Customers(
    Id INT PRIMARY KEY,
    CompanyName NVARCHAR(100) NOT NULL
)
GO

Create the first migration script from an existing database

If we work on a solution with an existing complex database we need a tool to create our first migration script. Tools are capable of tracking dependencies and relations inside the db schema to calculate a single script that evolves our db to the desired state. There are many tools on the market. In this example we’ll use Red-Gate SQL Compare. With this tool we select our existing db as the source (IC6.ParrishShoes) and our target to sync. As we explained before we need an empty database so we compare our db to that and we obtain our first script.

The script calculated by SQL Compare engine is the following.

/*
Run this script on:

        ..EmptyTarget    -  This database will be modified

to synchronize it with:

        ..IC6.ParrishShoes

You are recommended to back up your database before running this script

Script created by SQL Compare version 13.4.0.6563 from Red Gate Software Ltd at 03/08/2018 23:20:53

*/
SET NUMERIC_ROUNDABORT OFF
GO
SET ANSI_PADDING, ANSI_WARNINGS, CONCAT_NULL_YIELDS_NULL, ARITHABORT, QUOTED_IDENTIFIER, ANSI_NULLS ON
GO
SET XACT_ABORT ON
GO
SET TRANSACTION ISOLATION LEVEL Serializable
GO
BEGIN TRANSACTION
GO
IF @@ERROR  0 SET NOEXEC ON
GO
PRINT N'Creating schemas'
GO
CREATE SCHEMA [Ordering]
AUTHORIZATION [dbo]
GO
IF @@ERROR  0 SET NOEXEC ON
GO
PRINT N'Creating [Ordering].[Customers]'
GO
CREATE TABLE [Ordering].[Customers]
(
[Id] [int] NOT NULL,
[CompanyName] [nvarchar] (100) COLLATE Latin1_General_CI_AS NOT NULL
)
GO
IF @@ERROR  0 SET NOEXEC ON
GO
PRINT N'Creating primary key [PK__Customer__3214EC072CF93715] on [Ordering].[Customers]'
GO
ALTER TABLE [Ordering].[Customers] ADD CONSTRAINT [PK__Customer__3214EC072CF93715] PRIMARY KEY CLUSTERED  ([Id])
GO
IF @@ERROR  0 SET NOEXEC ON
GO
COMMIT TRANSACTION
GO
IF @@ERROR  0 SET NOEXEC ON
GO
-- This statement writes to the SQL Server Log so SQL Monitor can show this deployment.
IF HAS_PERMS_BY_NAME(N'sys.xp_logevent', N'OBJECT', N'EXECUTE') = 1
BEGIN
    DECLARE @databaseName AS nvarchar(2048), @eventMessage AS nvarchar(2048)
    SET @databaseName = REPLACE(REPLACE(DB_NAME(), N'\', N'\\'), N'"', N'\"')
    SET @eventMessage = N'Redgate SQL Compare: { "deployment": { "description": "Redgate SQL Compare deployed to ' + @databaseName + N'", "database": "' + @databaseName + N'" }}'
    EXECUTE sys.xp_logevent 55000, @eventMessage
END
GO
DECLARE @Success AS BIT
SET @Success = 1
SET NOEXEC OFF
IF (@Success = 1) PRINT 'The database update succeeded'
ELSE BEGIN
	IF @@TRANCOUNT > 0 ROLLBACK TRANSACTION
	PRINT 'The database update failed'
END
GO

Compile a database

If we run the previous script and we get zero errors we can say that we can compile our database. The same concept apply when we can execute a collection of migration scripts in the correct order without errors.

Being able to compile a database it’s a huge benefit because it enables us to create a database on demand by simply executing a list of script to get the desired version on which we can develop, run test, deploy to a new customer etc. We can also start to do some automation for our database tasks.

Automation with VSTS – Checking our migration scripts

An important part of having a collection of db migrations scripts is that we have to frequently check that this chain of scripts is valid.

We can invest in automation to do that and with VSTS is very simple to get started.

In the following example we create a build process that executes all the migrations scripts in our repository to a target database to check if everything can be completed without errors.

We go to the Build section of VSTS and create a new Build process from an empty process.

We browse in the marketplace the DbUp extension (free!).

DbUp is a .NET library that helps you to deploy changes to SQL Server databases. It tracks which SQL scripts have been run already, and runs the change scripts that are needed to get your database up to date.

2018-08-03_23-40-20

After the installation process of DbUb we can refresh our list of available task and add the DbUp Migration task. With this task we can execute scripts included in a folder of our repository to a target database. The configuration of the connection string is done with VSTS Build variables in the dedicated tab. We configure the script folder path and the Single Transaction strategy, too. This way all the scripts are executed in one single transaction and if we get an error everything will be rolled-back.

2018-08-04_00-06-31

2018-08-04_00-16-49.png

When we complete the configuration of the task we click Save And Queue.

The build process will start and if we’re a bit lucky everything will succeed.

2018-08-04_00-24-07.png

We can see that DbUp applied correctly the scripts by browing the db with SQL Mangament.

  • First 2018-08-04_00-07-34
  • After 2018-08-04_00-12-42

TL; DR

With this tutorial we’ve learned the basics of migrations scripts, how to create them with a tool and how to check if our collection of scripts is valid with automation implemented with VSTS and DbUp.

This is just the start

When we have such powerful tools we can start to think about automated processes that test our chain of scripts in many ways. We can do an “incremental” approach like the example above triggered at every change in the scripts folder. We can schedule a build process that every day at noon will execute all the migration scripts versus a temporary brand new database.

What will you do with this super powers? Let me know in the comments below!

Reference

VSTS DbUp Marketplace – https://marketplace.visualstudio.com/items?itemName=johanclasson.UpdateDatabaseWithDbUp

Red-Gate SQL Compare – https://www.red-gate.com/products/sql-development/sql-compare/index