One thing that I have learnt while working with the SQL Server CSS team is that there is always more than one way to skin a cat. It means that majority of the problems have more than one solution/workarounds but the feasibility of implementing/adopting would depend on a number of factors like business restrictions, environmental limitations, time (this is probably a limiting factor in every aspect of life) etc. Very rarely do we have a situation which doesn’t have a solution and results in a completely bizarre situation which starts reminding you of the puzzled expression on David Duchovny’s character in X-files.
So to arrive at the solution or I shall call it problem transformation into non-problem, we need to troubleshoot the issue which involves the following steps:
1. Asking the right questions
2. Collecting the right data
3. Analyzing the data and drawing intelligent corelations
4. Developing a hypothesis and proving the same with the data collected
Troubleshooting is indeed a scientific art. The more number of years you spend doing this, you will develop an acumen for it where seemingly ignorable data points would represent interesting patterns. And this is the art that we try and inculcate in everyone at Microsoft CSS. Understand this that we do NOT have crystal balls which shows us the problem on the environments that we are working on and nor do we have magic wands which solves all the problems on the server with a flick and a chant of abraca-dabra.
One of the limitations that a CSS engineer works against is that they have to understand the environment that you are managing as a part of your daily job within a few minutes of the call. This is what they are trained to do but we need to be cognizant of the fact we need to help them follow the set of steps mentioned above so that we can arrive at a resolution as early as possible.
I shall attempt to start a series of blog posts and create a roadmap for the most commonly encountered issues by SQL CSS teams. This would help you understand why the data is being collected and what the engineer intends to do with the data.
Click on the roadmaps below to get a list of action items that can be used to troubleshoot the issues mentioned below: