Site Blog

Update on Site Crashes

As I mentioned in the first blog entry, the Holonation website suffers from periodic crashes. Crashes occur for very different reasons. The reasons include:

• "We're sorry ..." -- That one is a bug in my code that I need to fix. There's less and less of those over time.
• "Rails app has failed to start properly ... " (or something lke that), is an artifact of the site being hosted right now on a shared server -- the single road into the shared server get congested sometmes (think rush-hour traffic in a big city)
  
• The site won't load at all.

The first two kinds of crashes are fixable. With the first kind, i can just fix the bug once I know about it. With the second kind, clicking the "Refresh" button will usually fix the problem.

But the third problem is death for the site. That happens when the shared server that this site is on get so congested, it either outright crashes, or comes to a virtual standstill (which is functionally equivalent to a server crash) (think 20-car pileup, freeway as a parking lot, people out of their cars craning their necks for a better look).

This week, for some reason, this #3 "death" problem has arisen just about every morning. I mean, for 5 or 6 days now in a row, but only in the morning.

The only good thing about this that I can think of is that usage this week of the site has been very quiet. By the members at least. Of course, that is another form of death -- "death by apathy" in that case, as opposed to "death by crash".

At some point, I'll need to review my hosting decisions for this site. That point will come when usage of the site starts to ramp.

But of course, we might never reach that point if this "death by crashing" continues every morning. Such is life trying to get a fledgling website going.

Anyway, hopefully this entry helps clarify your frustrations if you bump into these.