Website Hosting (Data Backup, and Redundancy) – a few tips

I was looking over one of my websites a few weeks ago on a Sunday evening, and I started getting major errors on the site. I logged into my server, and after finding nothing obviously wrong, I gave the server the command to reboot.

About 5 minutes later, after the site didn’t return, I realized that there was a real problem. I called my host and asked them to check my server out. They proceeded to reboot the server, several times, and again nothing. After several minutes of diagnostics, the cause was traced the the primary hard drive, which had failed, mechanically.

Now, I backup my sites regularly, and I backup my databases, but even so I wasn’t prepared for a mechanical failure of the server. I have a very good host, and they worked at 2AM to get my server back up by the next day. After about 6 hours they got my server back up with a new hard drive, reinstalled operating system, new control panel, and then they handed it over to me.

Even with backups of everything, it has taken several weeks to get all of the sites back up and running. The ecommerce websites were the highest priority, so they went back up first. Just Ftp’ing the data back to the server took over 24 hours because there was a lot on there and there were a lot of individual websites on there. Then there were a bunch of SSL certificates that needed to be re-created, the firewall needed to be re-configured. In all there were over 80 websites that went down with the server.

The point:
The point to this article is to remind website owners that they need to make very sure that they have a backup in the event that their website goes down. Nobody expects it to happen, and nobody wants it to happen, but it does happen.

My recommendation:
Most importantly, use a good web host. Whether they do backups for you or not, make sure that your server is monitored and manned 24 hours a day. If I didn’t have 24 hour support, I could have lost thousands of dollars in sales the next day because my website would not have been live.

Next, pay the extra money for a mirrored RAID setup with your server. Unfortunately, this server didn’t have RAID at the time, but RAID would have completely prevented this problem from becoming as bad as it did. The server would most likely have been down for a little while, but none of the data would have been lost. RAID alone would have saved hundreds of hours of work in getting all of the websites back up.

Make Daily or Weekly backups of all of your data, especially information in your database. Most websites are cached in Google which can make restoring static pages simple, but there is nothing that can restore your database. If you update your sites daily, back them up daily. If you update them weekly, then back them up weekly.

Don’t store backups on the same server as the websites. In the event of a mechanical failure, all of your data on the failed hard drive will be gone. Keep backups on a remote computer.

Note:
If you store sensitive information on your server, and you need to back it up, make sure that your backups are encrypted and not stored on a public computer or server.

Full on load balancing and redundancy:
For most websites this is a bit of overkill, but if you have the traffic volume, look into getting several servers. You can setup server clusters, use DNS balancing, or traditional load balancing. There are many ways to ensure website redundancy, so your website never goes down. These setups are very expensive, but if you have the need, then the price is probably not the issue.

If you are looking for a managed or dedicated server, I highly recommend checking out 800 hosting. They have reasonable prices and their support is excellent. Any company that drops what they are doing for 6 hours to build a server at 2AM on Monday morning, passes the good support test in my book.