Client health management is about as glamorous for your computers as oil changes are for your car, but they are both vital to the machine’s functioning. For the last couple months I have been immersed in a massive client health clean up project to get ready for a critical deployment using Microsoft System Center Configuration Manager (SCCM) 2007. During this project I have learned several lessons and compiled some best practices for keeping your clients healthy and fixing them when needed. These techniques are applicable for SCCM 2007, SCCM 2012, Altiris Deployment, and any other application that requires a minimum level of computer health to function.
Adopt a Three-Phased Approach to Client Health Management
The three areas of focus for client health management are provisioning, prevention, and a well managed break / fix strategy.
Provisioning
It is far easier to ensure a computer is healthy when it is installed than it is to fix problems after it is already in use. It goes without saying that a computer should work when it is first deployed into production. Client health issues can sometimes fly under the radar during this process. For instance, a computer with a BITS or WMI problems might seem just fine to the installer and the end user. These types of computer problems are likely to only become apparent when remotely deploying packages.
In larger organizations and even some medium and small businesses, the person installing the hardware may not necessarily have any experience with or knowledge of Config Manager. It is our jobs to communicate the necessary information so that when a computer is installed into production it is verified to be healthy.
Prevention
Prevention is the best medicine for the health of your computers as well. There are several applications, both commercial and open source, available to perform regular maintenance on your client computers. I don’t have a preference for what is used, but I recommend some type of scripted discovery of problems and automated resolution when possible.
Common problems that can be resolved programmatically include incorrect DNS entries, corrupt WMI, firewall problems, and SCCM Agent installation.
Managed Break / Fix Strategy
No matter how good your prevention discovery and resolution scripts are, there will always be exceptions. Without a managed break / fix strategy those exceptions will end up going round and round in your prevention cycle and never become healthy. A mechanism for catching those exceptions is critical so that they can be researched further. For instance a missing SCCM Boundary might mimic a problem with the SCCM Client. Likewise, a network connectivity problem might look similar to a Windows firewall misconfiguration.