I don’t demand 100% uptime from web hosts. I expect it, but I understand that things happen. So it took not one, not two, not three, but at least four site-down situations before I gave up on ANHosting / midPhase, the original host of The Schlog and The Schrug. And those were just the outages I know about because I happened to be trying to post or read something at the time. When I finally made the decision to seek out a new host, I found out just how bad the service was.
The most recent site outage was caused by a complete server crash — one that required ANHosting to move my site (and many others) to a new hard drive. When it came back up, I noticed that at least one recent Schlog post and the entire contents of The Schrug were missing. I contacted tech support to find out what was wrong. They said:
The backup of your account was done two weeks ago. Sorry, but the latest backup update failed, so the recent updates you made weren’t backed up
Thank you for your understanding
Are you friggin’ kidding me? You back up your servers every two weeks? And when a backup fails, you don’t immediately re-run it?
Fortunately, because I use an off-line editor (Windows Live Writer) to compose my blog posts, I had local copies of everything. (Hey, DR conference folks — do I get some props for that?) Still, I was more than a little ticked off, so just moving to another host wasn’t going to satisfy me.
ANHosting has a 99.9% uptime guarantee. It seemed at first blush that they had certainly failed to deliver on their promise. But as I thought about and researched it more, I realized that I didn’t really know what a 99.9% uptime guarantee really means to a web hosting company. I sent the following e-mail to their sales department (not letting on that I was already a customer):
You say you have a 99.9% uptime guarantee, yet your terms of service say you do not promise uninterrupted service. Can you explain precisely how your guarantee works? What exactly are you guaranteeing, and what is the remedy if you fail to deliver on your promise?
Yes, we do guarantee 99.9% uptime, the clause in the TOS covers that .1%
That said, things do happen, such as hardware failures, etc that can take your site down for more than .1% of the time and if that does happen your account will be credited based on the downtime that you experience, that is how the guarantee works. Before I finish sending you this E-mail, I thought I would write a little about who midPhase is and why you should host your website with us.
Then they added a bunch of information about their experience, awards, yadda yadda yadda.
I wrote back:
I should have been more specific. Uptime is guaranteed as 99.9% of what? A day? A month? A year? For example, if you say you will be up 99.9% of the year, you could still be down for an entire business day, as long as you were up the rest of the year. On the other hand, if you promise to be up 99.9% of every day, then that’s only 1.5 minutes of downtime allowed per day. At what interval are your statistics calculated?
And are you saying that the credit is based on the amount of downtime and on the standard monthly fee? For example, suppose my site is down for three hours, or 12.5% of one day. And let’s say that’s enough to trigger the guarantee provision. Do I understand that I would get a credit equal to the monthly fee times the percentage of the month that I was down? So that would be 12.5% of one day divided by 30 days times $6.95, or 2.8 cents. Is that right? If my site is down for three hours, you’ll give me a credit of 2.8 cents? And if my site is down for an entire day, I get a refund of 23 cents?
I love being snarky. By this time, though, they figured out that I was already a customer. So in an attempt to save my business, they wrote:
The uptime guarantee is measured on a monthly basis, and resets each month. I have applied a free month credit to your account which will automatically be applied to your next invoice.
The transfer process was a little trickier than I anticipated. I got some conflicting opinions from a couple different folks at HostICan tech support about how to do it. Options included using the WordPress export/import feature, an FTP transfer of the desired directories, and a third that I don’t recall. I tried a couple times and it didn’t work. Fortunately, HostICan tech support also gave me the option of letting them do all the work for me, and I took them up on it. I had to give them my username and password, but there was nothing at the davidschrag.com site at midPhase that was confidential, and I didn’t use the same username and password at any other site. So there was really no risk in letting HostICan handle it.
There was still a hiccup or two left after they transferred the data. In theory I should have been able to view the new site using an IP address before changing the DNS record hosting for davidschrag.com from midPhase to HostICan. But in practice I had to move the DNS records first and then refer back to the old midPhase site by IP address to pick up a few straggling bits of code.
At this point, though, I seem to be fully up and running at HostICan, and I thank Ted David at HostICan support for working with me on this over the weekend.
If you notice a problem with either blog, please let me know right away. It’s quite possible that something was lost in the shuffle.