When I got my present workstation last February (for the story, click here), It came with a single 1 TB disk. I planned to put two more in it, since there were supposed to be two free bays. When I opened up the case up to stick the new disks in, one of the bays was really hard to get to; in order to use it, I’d have to reroute almost all the internal wiring. I put a 1 TB drive in the easy-to-get-to bay, and buttoned the case up. I only had about 700 GB of images that I wanted to keep on the desktop, and I figured that, by the time I needed the space, I could put in a bigger disk.
Last week, I noticed that I had used more than 850 GB of space on the image disk, so I looked around on the web. There seemed to be several brands of 2 TB disks available, and I didn’t read many horror stories like those that accompanied the first 1.5 TB drives. I ordered a 2 TB Western Digital Green-Series disk, figuring that quiet, cool, operation and reliability trumped speed for me, since I wasn’t going to store any programs on this drive. The disk arrived in a couple of days, and today I installed it.
First I ran Vice Versa manually, just to make sure I had all the latest files backed up on the server. Then I disabled the script that backs up the Image disk. It wouldn’t do for the backup program to start running in the middle of the restore. Since I told Vice Versa to mirror the source to the target, having a backup run before the files were completely restored would cause data loss as the unrestored files were deleted from the server. Automation is a two-edged sword. It saves time and energy, and keeps you from making small mistakes, but it facilitates making huge ones.
I disconnected the computer and carried it to a workbench. Dell’s no-tools case was easy to open. I had the old drive out and the new one in in less than five minutes. Back upstairs, I connected the cables, started the computer, brought up the disk administration program, registered the disk and gave it a quick format. I prepared a Vice Versa script to restore the backup data to the new disk and turned it loose. It announced that it would be done in about thirty hours. That was fine with me; I’m going to have a new set of images to edit in two days. I set the old disk aside, thinking that I would take it to the bank to be an extra copy of my images.
Why bother writing about this, since it was no big deal and went smoothly?
The first reason is to note how accustomed we are to having advances in technology come along and save our bacon. I am reminded of a sign that I once saw on the wall of the VP Operations of a high-tech company. “Believe in miracles? We count on them!” it proclaimed. Disk data density has been rising at a once-amazing pace for so long that it is no longer remarkable. It hasn’t happened by accident. There are a few breakthroughs and many clever tweaks behind the seemingly-inexorable progress. We ought to stop for a second and thank all the engineers and the few scientists that are making it all possible.
The second reason is to remark upon how cheap disk technology has gotten. With spinning storage hanging in there at about a dime a gigabyte, I didn’t even think of paring down my image collection so that it and the next year’s additions would comfortably fit on my old disk. It wasn’t worth my time. Thank those engineers again.
The third is how long restores take now that the data sets have grown so large. Admittedly, I’m not doing the restore in the fastest way possible, just the easiest. I could have kludged both the old and the new disk temporarily into the workstation and done a disk-to-disk restore. The network restore allows me to use my workstation while it’s running, and also gives me an extra CRC check on every file to make sure there is no lost data.
The fourth is the mistake that I didn’t make. It would have been easy for me to forget that the backup script was running, and wipe out the backup copy. Since the script that copies the backup copy to the Drobo was running, I’d have lost that data as well. I would have had the original disk, but still, the lesson here is to think through the system operation before you act, and keep extra backups.