Voice of the DBA

Category: Blog

  • Git, GitHub, GitLab, BitBucket, What’s the difference?

    I was watching Brent’s webcast session on GitHub recently. I’ve struggled to explain this to users in the past, and it’s always good to see what others do when they do similar sessions.

    You can watch Brent’s session, and I’ve got some content on git. Need to do more for Redgate, so look for some content to appear on the Hub or our YouTube Evangelist playlist.

    In any case, what’s the deal with Git, et al?

    Git is a version control system (VCS). If you’ve ever used TFS for code, or Visual SourceSafe or Subversion, Perforce, CVS, etc., they are all methods of tracking versions of files. Usually for code, but could be for any files. Binaries don’t work as well, but any text, and plenty of people use a VCS for book projects.

    While git (or Git) is a VCS, there are various companies that provide git hosting services. GitHub is the most well known, now owned by Microsoft. This means that they provide essentially a remote file service for git projects. Since git is a distributed VCS, this means that you can have multiple copies of your git repository (folder) and sync back and forth. Like DropBox, but with programming code versions.

    BitBucket, GitLab, Beanstalk, Codebase, etc. are all hosting services. You could duplicate the git portion of hosting with a VM in Azure, AWS, etc. However, these services provide a few other things. They include some features like security, clients, and pull requests. A pull request is a method of notifying people to review your code before it merges, and it’s a fantastic team feature. It’s in git, but it’s cleaner with a service.

    I tend to use Azure DevOps for my git hosting, but I do have repos at GitHub as well. I’ve tried BitBucket, and really they all feel the same. I get used to a GUI or CLI push/pull of code and then it doesn’t matter.

    If you’ve never tried using git for code, give it a try today. I’ll get more content and videos up soon, so watch for those.

  • SQL in the City 2019 in London

    We’ve started setting up our SQL in the City Summits for 2019 and the first one is London on April 30. I’m going to start making my plans now and hope to see some of you there.

    We’re still finalizing agendas and getting some speakers lined up, but if you’re anywhere near London, take a day and come see us talk about DevOps, compliance, and Redgate.

  • Do You Have a SQL Server Estate?

    At Redgate, we do a lot of research. Maybe not as much as Microsoft Research, but we are always trying to investigate more about customer problems and challenges before we build or change software.

    Take our survey on SQL Server Configuration Management and you could win prizes. Or a prize.

    I’ve never thought about managing an estate, but not that I live on a horse ranch, that has a little more meaning. In the past, I’ve managed hundreds of instances and thousands of databases. That’s certainly an estate, but even when I had 5 servers, that is something of a small estate of things to keep me busy.

    At Redgate, we know there are lots of challenges in managing SQL Servers. We constantly have a team of people working on SQL Monitor, and they are releasing changes every week or two. In fact, we just released v9.0 this year, full of enhancements customers have asked for.

    In addition to the tactical team that’s building and fixing things day to day, we are doing future research into where to take our monitoring and tuning software. That’s some of what this survey is about. Let us know about which things might help you smooth the process of managing systems.

    Give us a few minutes and a little information and take our survey. You might find that a future version of SQL Monitor contains something you thought would be useful.

    Take the Survey today.

  • Don’t Leave the Database Behind…Or Win If You Do

    If you’ve had struggles with database development and had a real world disaster, you could win in a new Redgate contest. We’re looking for the places where old style, waterfall, traditional database development has been a problem. Especially when application software developers have been adopting DevOps, small, rapid release strategies.

    Enter before Mar 20 and give us your story.  Don’t worry, we’ll anonymize your name and company details. You can enter as many times as you like, but can only win once.

    My Struggles

    I started working with a small startup company that had 8-10 application developers. We tried to work in a rapid format, releasing a new version of our website every 2-3 weeks. This was going OK, but not great. Often we’d release and developers would go to test their new feature, only to realize that they’d forgotten a database component.

    They would roll forward, finding code and deploying a database change during the release process, usually getting the feature working in 10 or 15 minutes.

    Until they broke something else because they hadn’t really tested the database component complete. Or they realized they needed yet another database change. Or they had forgotten to code for some edge case that another developer tried during the release smoke test.

    When I started, we all gathered in the office on a Sunday night and spent an hour or two releasing code. Not one of my favorite times.

    The app developers were using source control, and trying to test, but they often sent code to QA and then fixed things in QA, forgetting that they needed to ensure those changes were deployed again to production.

    Across a few months, we implemented what people would not call a DevOps process, reducing our release time to a few minutes, with myself and another developer on the phone, releasing automated changes from a CLI.

    Database DevOps makes everyone happier.

    Share your story before Mar 20 and good luck.