Recovering From a Catastrophic Failure
My computer wouldn't turn on. I didn't know what had caused it. It would be a few days before I could track down the problem. Was I able to continue working?
The Problem
Something wasn't working. I don't know what. My computer rebooted itself several times, but I ignored the warning signs and kept on ploughing through - I was working with a program I knew to be unstable and I suspected it was causing all the problems.
Then it wouldn't boot up at all. Nothing. I suspected the power supply, but some of the lights on the motherboard were still on. The fan in the power supply refused to move when I tried to push it around with a pen. I knew I had to buy a new power supply, but what else?
Then, when switching the power supply on to test it one more time before I took it out, I heard a SNAP and smelled burning circuits - the power supply was definitely bad. I replaced the power supply with an Antec TruePower and it worked. Phew. But the next day I heard a BOING and my computer froze. I opened it up and the CPU heatsink/fan had popped off. Damned thing must have come loose when I was replacing the power supply. The CPU overheated and seized up. Useless. I had to get a new CPU.
All tolled, my computer was out of commission for close to a week. If I'd decided to get a new computer I would've been out of commission longer than that. I could have saved money ordering parts online, but again, that would've taken more time.
Worst case scenario: The PSU could have damaged the hard drives somehow, all of my information would have been gone.
So I used my girlfriend's computer for a few days. How did it go?
The Recovery
A few of the files I was working on were on the old computer. I could reconstruct them, but that would take time - it would be easier to fix the old computer and get them off of it. Other than that, I could have just kept chugging along.
While it may have violated my license agreement, I could have installed some of my software on her computer. I would have deleted it as soon as my old computer was working, so technically, I don't think I would have violated anything, but Macromedia's lawyers may say differently. I mean... the thing it was installed on wasn't a computer at the time I installed it on the other thing that actually was a computer so...
Most importantly, all of my most important files - the ones with the passwords to everything and all my registration codes for any shareware I may have downloaded were sitting on a server waiting for me to download.
All thanks to a little freeware program called SyncBack.
Setting SyncBack Up
First, I keep all of my most important files in a single directory called "E:\notes\." This directory includes some Treepad Safe files (Tranglos Keynote is a good freeware alternative) with information I've gathered over the years, including login information for websites, phone numbers and addresses, complete magazine articles I saw online and wanted to keep, chords to some songs I know and so on. I also keep some registration codes for programs, a few graphics files, some of my most important writings, and the database for my website.
If you're familiar with Getting Things Done, this is a big component of my trusted system.
Then I set up two profiles in SyncBack. The first copies files from one directory to another, zipping them in the process, and the second FTP's them to a server. You'll see that these profile are listed as "Not Scheduled." That's because I choose not to use the Scheduler within Windows to run these programs at a certain time, I use SyncBack's built in timer to run these scripts - as you can see, every 30 minute and every 8 hours.
The reason I copy the files to a second directory (E:\notes.zip\) is that they'll FTP faster and take up less space on the web server if they're zipped up first. Here's what that looks like. It's pretty simple to set up, and you can see from the interface, it gives you a step-by-step description of what will happen.
Here you can see that I compress the files using a conservative "Normal" compression scheme. For added safety, I can password protect them.
Then I set up a second profile to FTP the files to my server. As you can see, this is the same basic interface to FTP the files. The only difference is that the destination is the folder on the server that I want the files FTP'd to.
And here's the tab where I enter all my FTP information. Pretty simple stuff.
The directory is protected via .htaccess, and if I'm particularly paranoid, I can use SSL to connect to it.
Here's the tab where I set how frequently the script runs.
That's about all there is to it. Note that the "Email" tab is just so you can email the log files, you can't send the archived files to yourself via email. Sorry.
A Step by Step Overview of What Happens
- When I turn on my computer, SyncBack launches (It's an option under the Preferences menu).
- The files are copied from one folder to another, and it FTP's the files to the server (not necessarily in that order).
- Every 30 minutes any files that have changed are copied & zipped.
- Every 8 hours any of the zip files that have changed are FTP'd to the server.
- If my computer crashes, I can go to the website and retrieve my files.
- I check my mail via a web GUI that was set up beforehand.
- Most of my bookmarks (except my every-day ones) are in a Wiki.
- If I get my computer back up, I copy the new versions of the files to my E:\notes\ folder.
- I sleep easily at night.
As a bonus, I can access these files from anywhere in the world, including when I travel.
Update: March 10, 2005
The authors of the SyncBack liked my article so much, they've given me a copy of SyncBackSE. I gleefully upgraded. A lot of the features are transparent - better this, improved that, but what makes SyncBackSE well worth the money is the ability to create folders based on the date.
I can now create a rolling backup of the last 7 days by specifing the folder %%DAYOFWEEK%%, and each time it runs it will place the files in a folder based on the day of the week. Or I can create a rolling backup for the past 30 days by choosing %%DAY%%, or daily archives forever by choosing %%DATE%%.
This means I can have:
- Offsite backups.
- That are totally hands free and automated.
- That are secure using zip passwords and are encrypted in transit.
- Are as frequent as I want.
- Go back as far as I want in time.
This meets all of my criteria for a perfect Hands-Free Backup.
I can do daily backups from now until the end of time. I'm only limited by the space on my remote server, and with zip compression I'm not really worried about that either. You can learn more about variables on 2BrightSparks' website.
page first created on Tuesday, February 15, 2005
© Mark Wieczorek