back ups

summary

This subchapter looks at backing up systems and data.

While everyone should perform back ups, when a system crashes, everyone will turn to the system administrator to be rescued. And this is a get yourself fired kind of issue because the system administrator will always get the blame for lack of back ups.

back ups

This subchapter looks at backing up systems and data.

While everyone should perform back ups, when a system crashes, everyone will turn to the system administrator to be rescued. And this is a get yourself fired kind of issue because the system administrator will always get the blame for lack of back ups.

tools

Your basic tools for back ups are tar, various compression schemes, and cron scripts.

tar stands for Tape ARchive, from back in the days when back ups were put on those large tape machines you see in the old James Bond movies.

cron is a system for scheduling jobs, including UNIX shell scripts, to run at specific times.

rule of thumb

People are always asking how often they should do their backups, looking for a regualr time period.

The rule of thumb on back-ups is to back up your data before you reach the point where you have too much work to do to replace the data.

If you start typing an email and the system crashes, you are annoyed, but you can type the email over again. On the other hand, if you lose all of your emails, you have a problem.

Hence, the simple rule of thumb to back up as soon as you are about to have too much data to create over again.

Notice that this answer wasn’t in the form of a specific time periood (such as daily or weekly).

The reality of any large scale system is that you will be using the cruder method of specific times, such as daily, weekly, and monthly back ups.

incremental and full

A full back up stores everything on a system.

An incremental backup stores only those things that have changed.

Back ups (especially full backups) take a lot of storage space.

The typical policy is to do incremental backups fairly often (daily, hourly, whatever is appropriate for the demands of your system) and to do full backups less often (weekly or even monthly).

When you have a failure, you go back to the last full back up and then run every incremental back up. The more incremental back ups you have to run, the longer it will be before the syetem is restored, with angry users and angry bosses waiting impatiently.

three location storage

There is the rule of a minimum of three copies of your back ups. You have two copies of your back on site and one stored at a secure location far away.

The secure storage copy protects you from fires or eathquakes or floods or other disasters that physically destroy your ability to restore your system. Therefore, the secure back ups need to be kept far enough away that they won’t get destroyed in the same disaster.

Of course, this also means that there is travel time for using the secure back up.

The local back ups are usually alternated (or rotated, if you have more than just two).

The first full back up is stored on media A. The next full back up is stored on media B. The next full back up is stored on media A (wiping out the original back up).