Server installation and configuration

More than installing the operating system. It takes in kernel tuning, disk mirroring, reviewing and applying patches, hardening the security settings, and writing down what was done so somebody else can run it later.

By the real load, measured, not by the software vendor's generic recommendation. Sizing for an isolated peak means buying capacity that sits idle; sizing for the average leaves the environment slow at the moments that matter most. The way through is to measure how it behaves over time.

It means adjusting the limits and behavior of the operating system core for the load that particular server will take: open file handles, network parameters, memory management. The defaults are built for general use, not for a database or one specific application.

It means writing the data to more than one disk at the same time, so that one disk failing does not stop the server or lose anything. It is the most basic protection against a bad disk, and it does not replace backup, because the mirror copies an accidental deletion just as faithfully.

Security patches as soon as they have been validated in a test environment. The rest follow a scheduled window. The risk of not patching grows slowly over time, and the risk of patching without testing arrives all at once, so the answer is to have a process, not to pick one of the extremes.

The choice usually comes down to the application that will run on it, the licensing, and what the team knows how to operate. Integrity-UX works with both, so the decision can be made on what fits the case rather than on what the provider happens to know.

Because a server installed without a record turns into a black box: nobody knows why a parameter is set the way it is, and every change becomes a gamble. The documentation is what lets you rebuild the environment and hand the operation to somebody else.

It depends on the usage profile and on how far ahead you are looking. A steady, predictable load tends to favor your own environment; variable demand favors the cloud. In practice most companies end up with both, and what matters is deciding case by case, with numbers.

With an assessment of what the environment has to carry: which applications, at what load, and with what availability requirements. That is where the sizing, the standard and the installation plan come from.