So if you're a busy administrator who's been asked to manage SQL Server, what should you be paying attention to? What tasks should you be performing on a regular basis to ensure the health and well-being of your new charge? This simple list highlights the five most important things any administrator can do for their SQL Server installation, keeping it tuned, online and recoverable.
1: Do better backups
A solid disaster recovery plan – as well as granular data recovery plan – is your No. 1 tool for being a SQL superhero. Start using frequent transaction log backups in between your normal full and differential backups. Better yet, look for a “Backup 2.0” continuous data protection solution – vendors like AppAssure and Sunbelt Software, and others offer solutions – that will prevent you from losing more than a minute of data in the event of a failure. They also make restoring faster and easier.
2: Make a maintenance plan
While you’re backing up your databases, make sure you’re updating database statistics, reorganizing indexes and so forth. If your server is configured to start SQL Server Agent (a separate service) automatically, then use SQL Management Studio to create a new Maintenance Plan that’ll run automatically and accomplish many of these tasks for you. Don’t just set it and forget it, though;review the plan’s job log periodically to make sure it’s working.
3: Back up the master database
You might not realize it, but master is where SQL Server stores all of its own configuration settings. In a disaster recovery situation, you need a recent backup of master to bring your server fully back to life. You don’t necessarily need minute-by-minute backups, because master doesn’t change that often. However, any time you make a server-wide configuration change, or add or remove a database, update your backup of master.
4: Configure operators and alerts
Within SQL Server, provided SQL Server Agent is running, you can configure operators that represent real-life human beings. Provide your work schedule and contact information to SQL Server and then configure alerts, which are situations that will generate a message to you – if it occurs during your specified work schedule. This is a good way to keep on top of database files that unexpectedly run out of room, server-wide issues and so on.
5: Tune your indexes
Aside from rebuilding and reorganizing them on a regular basis, tuning your indexes – something I do once every year, year and a half on a busy database – gives SQL a chance to re-evaluate your database given its current condition and workload. Use SQL Profiler to capture real-world database traffic and feed that information to the Database Engine Tuning Advisor to get solid recommendations on changes that can improve performance. Today’s index design is not necessarily the best design for tomorrow, so occasionally revisiting the Advisor is a good idea.
See, nothing difficult. No tools you have to buy, and nothing you really have to add to your existing daily workload. Just a few simple tasks will keep your databases running more smoothly and make sure you’re covered in the event of a failure.