Over the past year, we’ve gone through these issues with GoDaddy, BlueHost, Typepad and many others. One goes down, tons of sites go down. But it doesn’t have to paralyze you. A little planning can make it easy for you to flip the switch to hold you and your site visitors over.

Sometimes bad things happen to good servers and suddenly a chunk of your online presence is down with the dreaded 500 error. We’ve seen it with GoDaddy, BlueHost and many others. It can happen when your blog gets hacked – WP and the like, when your cart goes wonky due to SSL certs…. possibilities are endless. Do not worry. I’m here to help you prepare for “those times.”

DNS – Domain Name Servers – tells the world where to find the parts of your site and presence:

  • mail.yourdomain.com,
  • blog.yourdomain.com
  • podcast.yourdomai.com

You get the picture.

Make a note of the IP address for EACH of your sites. Use CMD Prompt or other tools/sites to find out. 
Consider setting up free hosting accounts on a few UNRELATED suppliers, just in case. You may already have this in place if you blog, have a cart, a site, podcast, etc.

You have a main site, right? It has an FTP file structure (to upload your files), right?[slogan]*if at this point you are glazing over, show this to your IT and/or your web goddess to have them help you with this.*[/slogan]

Back to the plan: For EACH of these types of landing pages/subsites, I want you to create a SINGLE page with the branding for that subsite. Include: contact information – email, phone, subscribe, dynamic twitter feed so you can easily post updates to the status of your broken site, a humorous image to add some levity and a sincere message.
I also suggest you have a contact form that saves data to a database that you can maintain from Google Drive and just have it in there to hold you over in case your database stuff is down. You can choose to use it or not. At least you won’t miss a lead with the form submissions and won’t have to retype it because it will be saved. Acknowledge that you know the form is REALLY ugly but it will get the messsage to you. Google forms are hideous.

MAKE ONE FOR EACH. Have an images directory in this new subdirectory for each with the images related to THAT landing page contained there so you can use relative paths (this will make it easy to upload the same folder to all of the other servers in case you need it.

Upload these packages to the FTP root level of EACH of the other subsites. You never know which one will be the only one working.

Now, go to where these domains are hosted and add a hosting account under each of them for these subsites.
Yes, for each. You don’t have to change anything yet, just have it ready to flip a switch. Have it point to that subdirectory you created. It will look for index.htm, index.html, index.php, index.asp in that order usually. Make sure you name the one page index.html or index.htm to keep it easy. You only need one page and the related images folder in the directory.

Now you are ready.. LET’S HAVE A DRILL – ready to practice? Here we go…
“OH CRUD! MY SITE IS DOWN AND I HAVE A SPECIAL ONLINE EVENT TODAY – CRUD, CRUD, CRUD!”
Don’t panic.
Check your other sites – which ones work? Which are the easiest to work with?

Use cmdprompt or other tools to ping the site that IS working: command would be: ping www.yourgoodsite.com
You’ll see the IP address. Make a note of it.

WHERE IS YOUR BROKEN SITE DOMAIN HOSTED?
LOGIN.
Go to the DNS for that domain.
Change the @ record under the HOST RECORDS to point to the IP address. Change update TTL time to 6000 (10 minutes).
Update.
That will change it, invoke the “stand by” hosting account and direct people to THAT page.

Keep testing the broken server and hosting company status to see when it is fixed. Usually they post updates on Twitter. Typepad went through this recently – this is how I came up with this solution. Really helped my client.

Once everything is really fixed and stable again, change your IP address BACK to the one that was broken for the @ record in the domain. Leave the hosting account “on call” and wait for it to propagate back to how it was before it broke.