One of my first jobs in the computer industry in the mid-90s was that of a Computer Operator. Now it sounds like a funny job title since, after all, who doesn’t operate a computer nowadays? It was, and still is, a specific role in a large production environment where mainframe computer systems and industrial-grade computers are still in use.
As a Computer Operator for the Geological Society of America, I would have to know how to process and retrieve data from files and databases in a variety of media, including 9-track tape devices, the kind you see in 1950’s science fiction movies and old IBM documentaries depicting computers.
There I learned the three most important jobs of computer operations which are, in order of importance:
Backups, backups, and backups.
Like most things boring and mundane — exercise, doing yard work, and taking vitamins, to name a few — performing regular backups at least weekly, and as important, verifying the backups, can keep you out of trouble. You don’t need to be a company as big as Google to see the necessity of this task. Human error and data corruption happen all the time and while the time spent rebuilding a computer system is not trivial, it is fixable. Your precious data such as photos, documents, and scans, on the other hand, may not be recoverable without good backups. The best way to ensure that data is available in the event of a catastrophe is to have a whole copy of it.
One theory behind “Cloud Computing” is that multiple copies of your data exist in various locations at once and are continuously being updated. While this sounds cool in theory, the reality is copying data like that is not only less secure, making more copies vulnerable to hackers, but there is strong likelihood your data will fall out of sync at some point. I wouldn’t rely on any single technology any more than I’d rely on any single individual or entity when it comes to backups.
There is the old chestnut story about the backup operator who was diligently making the same backup on the same backup tape week after week. The data had been deleted by someone two weeks prior. So, of course having a backup from one week ago didn’t help.
Ideally you will do a complete backup of your critical files at least once a week and rotate them out to one or more offsite locations every week. Having four different drives is not uncommon, as well has keeping an annual or semi-annual backup just to be safe. External USB drives are ideal for this backup method. Once you establish a system test it periodically and always check the log files to ensure what you think happened actually did happen.
“Mission critical” computer systems are often built on a RAID system. RAID stands for “Redundant Array of Inexpensive Disks” (depending on who you talk to). This means your data and/or operating system is spanned across two or more disks so that if one hard disk fails, the others are able to continue to work to access the data. This is a type of fault tolerance to ensure systems stay running and data remains available, but it is no substitute for having data backups.
We had at one point a web system that was running a RAID supporting both the OS and the data, but when one of the hard disks became corrupted it simply populated that corruption to the rest of the system. Luckily the system was able to limp along while we rebuilt a new system from scratch and transferred the good data over to it from backups. The system was up and running at full capacity again in a matter of hours.
Like most teenagers supposedly don’t learn respect for the destructive power of an automobile until they impact something, so most computer users don’t respect the power of having backups on hand until there’s a problem and their data is lost. They are too busy and important to be bothered. Magical thinking. Don’t be that guy.
Also, don’t rely on anyone else to have backups of your data. Your IT person or webhost may promise they are backing up your data, but if something goes wrong and they didn’t do it, excuses will matter little. You having access to your website and computer data at-the-ready is really the only solution. As the aphorism goes, “If you want something done right, you’ve got to do it yourself!”.
And always remember the three most important things in computer operations: Backups, backups, and backups.