One afternoon on Apr 13 (even though it was Saturday 13, not Friday 13) all of my sites went down. I soon heard the terrible news… the server had crashed and wouldn’t reboot. That night Murphy’s Law came into effect, and the server remained down until a new and empty drive was installed the morning after. From that morning on, I was told I was on my own. The following is what I learned from this experience.
1. Back ups are lifesavers, if done correctly
I had fairly recent or full back ups (on my PC) of all the sites I’ve designed, a handful of which are not my own.
I did need to back track and redo a few recent changes, like content deletions I had made in the weeks prior.
There were glitches, despite all of my precautions. For example, I had depended solely on CPANEL for my forum database backup. Apparently the database had grown so large that when downloading the daily backup, a part of the database had been cut off. Large pieces of a forum database were lost.
Now I know to back up in multiple forms, like:
a) Structure only and data only.
Tip: Prune the data regularly to avoid mega-sizing your databases.
b) Compressed and not compressed.
c) Software/browser back up and CPANEL back up.
d) Table-specific back ups on larger databases (especially for USERS table!)
e) Saved files (incl. install files) for software and scripts that you use, including their MySQL and Admin username/passwords.
f) Full site back ups and home directory back ups.
Since the crash I keep all the above back ups, almost in an obsessive compulsive way. It may sound like a bit much, but you can create a back up schedule that’s near foolproof and bandwidth frugal. You won’t need to back up in every single way, every single day. Each site will have different needs when it comes to backups.
2. Not everyone backs up!?!
It’s true that webhosts back up their drives, usually every night. However, in our server nightmare the back up was corrupted by an employee of a company they outsource to. He ended up shattering that server-backup, which a few webmasters had been counting on. There were probably dozens if not hundreds of webmasters on the same server, and that unfortunate few had neglected backing up their clients websites.
As long as you’re planning on having a website presence, it’s almost guaranteed that at some point you will have a server crash. Get in the “safe keeping” habit, before it’s too late and you end up missing your site(s), or having forty angry clients calling you and demanding their (forever lost!) websites be “up and running pronto!”
3. Spread your sites out on more than one webhost
This is a great advice, that would have helped me tremendously. It would have been a way less stressful if only half of the sites I manage went down with that crash. I will listen to this advice from now on.
4. When it happens, stay calm
Our websites were down for 12-24 hours, and in a few cases 36 hours as I moved some sites to a different host. I worried about my regular site visitors, Google ranking, Adsense, and the long haul of getting it all pieced together again.
– My ranking didn’t budge at all, and spare the off chance you crash during a major Google update and can’t get your site back online before it’s done, yours will be OK too.
– My AdSense revenue did go down, but it wasn’t a total bust. The month before the crash I had an increase in AdSense revenue, and after the crash it had reverted back to normal.
The best thing you can do when your server crashes is stay calm. Likely, despite your best efforts, you will need to tweak programs here and there and put that thinking cap to good use.
I had to step away from the computer for a couple hours to catch my breath, afterwards I started reconstructing all the sites. Oddly enough, the crash wasn’t the most stressful thing that happened that month.
The crash, or more so the actions I had to take afterward, actually increased my confidence in my own abilities. I had a list of items to restore and/or fix, and with each cross out my mood improved. I gained the knowledge to make sure a system restore would flow easier if it happens again. The day after the crash ended up being a highlight of that month. The dreaded “server crash” wasn’t so dreadful after all.