First, a little history. Mac360 started life in early 2004 and was launched on a sunflower iMac that sat atop my home office desk. The Mac ran OS X Panther Server. Although there was not much horsepower in that old Mac, and bandwidth for OS X Server was limited, a popular Mac blog was born.
Hello, Goodbye, Hello, Goodbye
Through the years since, Mac360 was served from my home office on a Mac mini, a Power Mac G3, then a more robust Power Mac G4. Uptime for the Mac back in those days easily exceeded 99.9-percent, which is decent considering that the Mac and OS X are not great web servers. As visitors to Mac360 increased, I moved the site, and dozens of other sites that I manage, to a variety of server hosts, finally settling on a smaller server host a few years ago, one with a very good reputation for reliability and support; A Small Orange (ASO).
ASO provided all of our sites with three virtual private servers (VPS) and despite the increase in monthly costs we were pleased with both support and performance for years. Often the Linux-based servers would exceed 180-days to nearly 300-days of uptime between reboots. But all was not so smooth sailing. ASO was a boutique server host, home to tens of thousands of websites, and while that number sounds high, it was small compared to giant Endurance International Group (EIG) which bought A Small Orange a few years ago. That’s where the journey detoured and where my big bang theory comes in. We’ve experienced the big bang and it was not big fun.
Enter The Borg, Laughing
Over the past year, Mac360, and thousands of others websites, have experienced a growing number of server and network outages; not minutes, not hours; but days in length. The latest started on Christmas and was partially resolved just today, nearly four days after it began. Read this missive and you’ll get an idea of what EIG does to web server hosts when they are purchased by what some call Borg. They are assimilated and they die. For four days ASO was dead to thousands of websites thanks to network and server issue which could not be resolved in a timely manner (and still has yet to be completely resolved as you’ll note here and here).
The thing to remember is that web server outages occur everywhere, can occur at any time, and the duration can be as little as a few minutes, but may be a few hours depending upon the nature of the problem that caused the outage. Outages that last for days are far less common, but seem to be the modus operandi of large facilities and rapidly growing server companies who focus on growing a business by buying competitors rather than by growing organically and caring for customer needs.
EIG is a huge operation with database and server host facilities all over the world. You may recognize a few of their more prominent server host brands; HostGator, BlueHost, HostMonster, JustHost, Arvixe, Host9, Site5, and many, many others listed here.
While Mac360 and a number of the other websites I manage are back online, over one third of our ASO-based sites remain inaccessible. Customer support during this outage, now nearly five days in length, has been limited to a few canned phrases on a status update page every few hours, and an email daily with the same message. Phone calls were not answered and support would not or could not respond. Needless to say, we are in the process of tracking down a more reliable web server host, and there are many from which to choose. What there are few of, it seems, are good, reliable, dependable organizations. Any suggestions?
There are a few morals to this story. 1) don’t put all your eggs into one basket, 2) pay attention to the trends and respond accordingly, 3) less can be more, but bigger isn’t always better, and 4) backup, backup, backup (we have multiple backups on disparate server locations). Oh, one more. Remember the knight that stood guard over the Holy Grail in Indiana Jones and the Last Crusade? ‘Choose wisely.’