Category Archives: Internet

Web hosting adventures

While A2Hosting has a pretty darn good bang for the buck, I recently had to look for a web host with even more advanced features. For instance, I needed to be able to compile a more recent version of the git DVCS, and use it on my account of course.

So I looked around some more and settled on WebFaction. I had wanted to try them out for a while now based on all the good reviews you can find on the Internet on them, and now that I have I must say that I’m pretty impressed.

I was able to compile git using my SSH account, along with the dependencies to be able to use git-svn, and it all went pretty flawlessly.

The control panel at WebFaction is awesome, easy to use, and it seems one can host just about everything on their servers: anything PHP, all sorts of Python and Ruby frameworks, it just looks fantastic.

In most cases you are running your own private Apache daemon, so you control your own configuration, add the modules you want, etc.

Pretty good stuff, check them out!


A2Hosting makes another happy customer

I’ve been hosting this blog over at A2Hosting for a while now. I was looking for a reliable, developer-friendly, not too expensive web host, and I must say I’m pretty satisfied so far. It seems that they are providing just the right balance between price, flexibility, and performance.

I found a responsive and knowledgeable customer service that has always proved helpful, their servers’ performance and connectivity has remained top-notch to this day, and their referral program is quite nice, which doesn’t hurt.

So all in all they well deserve the thumbs-up that I’m giving them right now.

Thanks A2Hosting!


Fun with Apache’s mod_rewrite

Part of my duties over at Cobi involve taking care of some web applications we built for our clients. We based those on top of Linux, Apache, MySQL and PHP (the so-called LAMP platform.)

We use Apache’s deflate module to automatically compress the sometimes huge web pages that get sent to the browsers in order to save bandwidth, and gain speed.

Then the need arose to leave some of the pages uncompressed.

It looked quite easy, all is needed after all is to set the Apache’s internal variable no-gzip each time we need to skip the compression step.

Because the URI’s for those uncompressed requests do not follow a particular pattern, I thought that adding a “no-gzip” parameter to the query string and configuring Apache to set (with the help of the module setenvif) the internal variable no-gzip wherever it would encounter it would be enough.

But I was wrong. You can’t do that. Well, at least not with mod_setenvif, because you can’t use the query string with the SetEnvIf directive.

Take #2, I devised that I would then put my “no-gzip” string inside the request URI itself, even if it smelled like a kludge, and then SetEnvIf would be able to see it and do its job.

But it didn’t work either! Because, we happen to also use the rewrite module, with which we rewrite all URI’s so that the web application’s FrontController gets launched for each request. And it seems that mod_rewrite gets to rewrite the request URI before mod_setenvif gets to touch it.

So mod_setenvif was always seeing the FrontController’s URI, without my little “no-gzip” addition, and compression could not be disabled.

So back to my first idea with a “no-gzip” at the beginning of the query string, but with a twist: it is now mod_rewrite, not mod_setenvif which is setting Apache’s no-gzip internal variable:

RewriteCond %{QUERY_STRING} ^no-gzip
RewriteRule ^(.*)$ $1 [NS,E=no-gzip:1,E=dont-vary:1]

Note: it didn’t work until the variable got an actual value assigned. In other words, E=no-gzip didn’t work, but E=no-gzip:1 did.


How DNS works – part 1

Even though the DNS (Domain Name System) protocol isn’t hard to understand, there seems to be quite some confusion about it.

So I’m going to explain a little what it is, why we need it, and how it works. Hopefully I’m going to make it short.

What is DNS?

Roughly speaking, the domain name system is like a phone book, but for computers. It is a big directory for computers to find other computers.

Why do we need DNS?

Why would a computer need a directory to locate another computer, as they are all connected already through the internet, aren’t they?

Well, computers connect to each other in a way that is a bit analogous to the way telephones connect to each other. When making a phone call, you know who you want to talk to, their name, maybe where they physically are located, but you need to dial the right number to reach their phone, because those devices only work with numbers, not names or anything else. In the same way, if I type in the address bar of my browser, my computer doesn’t really know to which one, among the millions (billions?) of other computers on the internet, it should connect to. It needs to have the numerical address of Google’s server, what is called its IP address.

When you came to read this particular blog post, your computer took the web address (called a URL) that was mentioned in the link you clicked, queried a DNS server to get the only piece of information that it can use to connect to the machine that hosts this blog, which is: the IP address for

So the domain name system is about names, and their correspondence to numbers. It makes it easy for humans, who won’t remember those pesky numbers unless forced to.

It also eases the job of system administrators, who mention domain names in the configuration of the machines which need to connect to other ones, so that when one server gets moved to another network, they don’t have to reconfigure everything, they just put the new IP address in the right DNS server, and that’s it, no need to modify the configuration of the other machines.

How does DNS work?

First, let’s take a quick look at the anatomy of the names used with The DNS protocol, taking “” as an example.

The whole thing is called a hostname. It is a name that refers to the IP address of a computer (the host for Google’s search page.) This hostname is composed of several parts, separated by dots. Technically those parts are called labels, and in our example there are three of them: “com”, “google”, and “www”.

Notice that I listed the labels in reverse order, from right to left.

That may seem counter-intuitive, but those parts represent a cascading, tree-like hierarchy of domains and subdomains which is read from right to left, with the topmost domain being the rightmost one: “com”. This topmost (rightmost) domain has a special name: it is called the TLD (Top Level Domain.)

I said the domain name system is like a phone book, but in reality, it is a huge, hierarchical collection of phone books. This is important, as when the time comes to change DNS records (which are contained in a DNS zone, served by a DNS server) one must find out who’s the authority for that domain, because this party is the only one with the power to do so.

So, for the sake of keeping this blog post short, let’s see right away how the act by which a computer will retrieve the IP address corresponding to a hostname, the act of resolving the hostname by performing a DNS query, is done. The resolver will ask a DNS server for the IP address corresponding to the hostname, wait for an answer, and behind the scenes, the DNS server will itself perform as many DNS queries as needed to get the answer, traversing the DNS hierarchy in this manner:

  1. First, it has to start somewhere, it will query the root DNS servers. All DNS servers must know in advance either the IP addresses of those root servers, or the IP address of another DNS server which know either the root servers, or another DNS server which does, etc. Otherwise it can’t work. The root servers know who is responsible for each TLD. So our DNS server picks one of the root servers and asks: “what is the IP address for
  2. The root server answers: “I don’t know. But I know who’s responsible for the “com” TLD, here are the names and IP addresses of their DNS servers. Pick one.”
  3. So our DNS server proceeds to send the same query to one of those servers: “what is the IP address for
  4. The DNS server responsible for the “com” TLD then answers: “I don’t know. But I know who’s responsible for the “” domain, here are the names and IP addresses of their DNS servers.”
  5. Our DNS server then sends the exact same query to one of those DNS servers at Google: “what is the IP address for
  6. Finally, an answer in given. Google’s DNS server says: “Here’s the IP addresses for, pick any one.”

When a DNS server hasn’t got the answer, it can proceed to query directly the next DNS server in the hierarchy, if the resolver asks for it. A DNS server with such capabilities is called a recursive name server.


I hope this summary information is helpful. I think it goes a long way explaining the basics of the domain name system.

I will post a second part in a few days; or weeks, who knows. It will cover domain registration, WHOIS information and others.