Documentation – this is the main way to avoid repeating mistakes and to use that experience. Rules: Ask the administrator careful documentation and diaries of their actions, arguing in detail and describe in detail the working protsessy.Predostavte his team of software for general documentation (eg, Wiki), that these documents can be easily found by Emergency situatsii.Obespechte administrators disciplined adherence to document and periodically check them, for example, by asking questions such as: "When this was done, what script, anyone? What did you do in such a day? and so on. " If they were not able to answer confidently, so again they have ignored the documentation, trusting only his own memory. Code.org addresses the importance of the matter here. "We do not need to monitor the system. Users always notify us when something becomes not work.
" When users report a database administrator that something is not working properly or not work at all, it is often too late. Not control the system – means not being able to predict its future states and to influence them. Rules: Get a good monitoring system, allowing pre-identify and resolve potential problems before they lead to failure and downtime database dannyh.Staraytes avoid any problems due to system upgrade, actively working with developers and testers to ensure that new software will work without oshibok.Izuchite system performance its normal work, to quickly recognize the problem when it happens. When something goes wrong, first of all ask yourself the question: "How did it work?" And then "What's changed?", It will quickly realize that proizoshlo.Sobirayte statistical information on the system.