Welcome back to The Ironic DBA Files, a series where the newbie DBA on staff at Dallas DBAs chronicles the ups and downs of fast-tracking a new career as a DBA.
In the previous post I rambled about for a bit talking about primary keys and clustered index keys. That writeup was very tip-of-the-iceberg stuff and there’s much, much more ground to cover on that topic in the future. This week I circle back around again to touch on a topic I’ve already written about in the past: Maintenance Plans.
You’re Gonna Use That?
I’ve learned that Maintenance Plans have something of a bad rap among experienced DBAs, and I can partially see why. If you’re new to this game like me, let me give you the two main reasons I’ve pinpointed as to why a Senior DBA might tell you to not use Maintenance Plans.
- Misconfiguration: Like so many others functions within SQL Server, it’s very easy for an Accidental DBA to choose the wrong options and parameters. I’m not going to run through best practices here, but simply knowing that you can screw up your database by incorrectly implementing a Maintenance Plan is something to be aware of.
- Inferior Product: While SQL Server is itself an enterprise-level piece of software powerful enough to encourage companies and governments worldwide to use it for their data, let’s admit that it’s not the best at everything. When it comes to performing regular maintenance tasks on your server, there are better options out there if you know where to look.
Here’s the core issue with both of the complaints above—an inexperienced DBA doesn’t know any better. And as we’ve discussed before, so many DBAs are not trained, let alone full time. They were thrown into the server room by the boss and told, “fix it!” just because they were the staff computer geek.
If an Accidental DBA digs just a little bit into the workings of SQL Server they’ll discover the Maintenance Plan Wizard and think, “Hey! This is easy!” The problem is that the Wizard assumes you know what every available option does and how it affects your data. They also don’t have the time or experience necessary to understand how to properly implement a third-party set of scripts. Think about it, if a new DBA doesn’t understand how to properly implement a built-in set of maintenance scripts, how can they be expected to implement a more complex third-party product?
Like me, a new DBA is going to look at the Maintenance Plan Wizard their first time out and see the shrink option and mistake it for compression. Compression the database is a good thing, right? So shrinking my database is a great idea! Making a mistake like this can create a database that performs horribly, and the Wizard makes no attempt to prevent you from making poor choices. Even newer versions of SQL Server with an updated Maintenance Plan Wizard don’t help a junior or apprentice DBA understand what the options mean.
Brent Ozar (b|t) said, “If maintenance plans are designed for the accidental and junior DBAs amongst us, I don’t think a lot of this stuff [options included in the Maintenance Plan Wizard] should even be an option. It should just default to the right thing, and take care of the database with Microsoft’s best practices set up as standard.” SQL Server 2016 Maintenance Plan Improvements.
So here’s my attempt to help my fellow apprentice DBAs better understand how to implement a rational set of Maintenance Plans that won’t simply end up making things worse.
Maintenance Plans for Apprentice DBAs
First of all, the best thing you can do for yourself is to go and learn what the various options in the Maintenance Plan Wizard can do, and what implementation of those options will do to your database.
Second, one of the very first things I learned from Kevin is that you should separate all the different Maintenance Plan options you intend to implement into separate tasks. Simply put, while it’s easy within the Wizard to throw together one plan that does an integrity check, an index rebuild, and a full backup, it’s always better to keep them separate and scheduled during appropriate maintenance windows.
Third, related specifically to backups, another guideline Kevin shared early on is that any server in production should be using the FULL recovery model. Why? Because otherwise you can’t perform Transaction Log backups, which means you lose the ability to apply point-in-time restores.
If you don’t know this already, a log backup truncates the transaction log when using either the full or bulk-logged recovery models. A regular log backup helps keep your transaction log from growing uncontrollably, allows you to perform the aforementioned point-in-time restores, and limits data loss during database recovery.
Other quick points:
- Perform Check Database Integrity Often: Do this DAILY if possible, but at least WEEKLY. Always schedule this during a wide maintenance window if possible. (Read more about CheckDB back in Episode 2.)
- Learn the Difference Between Rebuild and Reorganize Index: It sounds simple, but this is an instance where you need to understand a bit about what’s going under the hood of SQL Server and how rebuild and reorganize are not the same thing. One of the chief points to understand and learn more about is the fact that rebuilding indexes automatically triggers updating of statistics, while reorganization does not.
- Create and Schedule Backups to Meet RPO and RTO Standards: If you’re working on a server in production, don’t just throw together a set of backups and call it a day. You need to know what your company/client expects in terms of acceptable data loss and down time. Your set of Maintenance Plan backups must be crafted to fit these expectations.
- Never Use Shrink Database: NEVER. Perform this manually on demand when necessary. Check out Paul Randal’s (b|t) in-depth look at why you should not shrink your data files.
- Learn About the Other Tasks: There are three other major tasks you can perform with a Maintenance Plan—Maintenance Cleanup, Execute SQL Server Agent Jobs, and History Cleanup. I won’t cover these here, but go and learn about them once you’ve got a handle on the other steps above.
That’s all for this week. Join me next time for The Ironic DBA Files—Episode 6: Return of the TSQL.
The Ironic DBA Files
- Prequel: The Ironic DBA—Starting a New and Unexpected Career
- Episode 1: You Back That Up?
- Episode 2: Attack of the Corruption
- Episode 3: Revenge of the Index
- Episode 4: A New Primary Key