Back in the days of running Novell, I remember the administrators comparing their skills by the amount of time a server could be online without restarting. Thirty days was a good time. Some guys even ran as long as six months. This was a badge of pride, and even an edge on a resume if they needed to look for another employer. Novell was primarily a host for shared files, although it could host many other things as well.
By comparison, today we tend to host our files and services on Linux or Microsoft systems for the most part. Moreover, we tend to isolate different services on separate virtual machines or physical machines, allowing for isolation of utilization patterns. Our expectations have increased to a point where six months is considered nothing special when it comes to up time. Out of the box, a modern day operating system is quite well managed, and tends to be restarted only when an update requires it. Even then, the Operating Systems are getting better at managing processes differently so they don’t always require a system restart to introduce change.
What I’m wondering is how well we take care of our database services and their hosts in our contemporary world. Hardware usually fails early if it is going to fail, and often lasts until it’s time to retire a machine for a newer model. The software tends to be more self-managed. In many cases you can stand up a smaller departmental size database with a smaller number of concurrent users, say under 20, and simply turn it on. All you really need to do is make sure you have a disaster recovery plan and you’re good to go.
It seems to me that a simple process is attractive and is often the implementation. Even using cloud based services, I would think there is little effort given to maintaining your database services. Granted, there is less you need to do in a Software as a Service implementation than self-hosted, or Platform as a Service.
Tomorrow I thought I would start a thread on a list of minimum actions one should do when maintaining an instance of SQL Server, even for a small departmental database. Today is your opportunity to share your thoughts. Why not leave a comment with your favorite management tasks you find are essential to the ongoing health of your server and services? I’m sure many of you have things you have learned through the school of hard knocks.
Cheers,
Ben