I found these while going through stuff in my documents, I think that they are good general guidelines
- Carefully research and document problems before making changes.
- Don’t make changes for the sake of change.
- Have others sanity check your proposal before implementation, then test and retest the changes.
- Explain the risks to management and IT colleagues before deployment.
- Maintain good backups, and develop sound restore procedures.
- Keep detailed configuration information for all systems.
- Have a tested local Administrator username and password for all systems in case you lose any network connectivity.
- Distribute a step-by-step procedure before rolling out significant changes.
- Develop a what-if list of things that could go wrong and the appropriate fallback procedures. This list should include plans to address potential problems that might occur days later (e.g., a server failure).
- Notify users well in advance of when a project will begin and end and of any planned outages.
- Keep a written log of activities during the changes.
- Carefully review logs to monitor system operation before and after implementing changes.
- Test the results from a user’s point of view. Use a non-privileged user account from a typical workstation.
- Coordinate with support staff and try to anticipate any user education needs that might arise.
- Document the procedures and save these files for future use.
Here are two related sysadmin books that I find helpful…