What makes nginx better apart from speed and resource handling?

It's not just a webserver, its also a really clever proxy server. You can use it to proxy anything (not just http, but any kind of tcp), from a frontend to a backend service really easily.

More importantly, it is asynchronous / event driven using 1 thread per CPU (core) , rather than using multiple threads/processes to handle responding to individual clients like Apache does, so it scales a lot better as well as being quicker.

Its not difficult to learn either, despite the nay sayers. The configuration format is different, but easier to learn than Apache's.

Good explanation. I currently use nginx on my RPis where required and Apache on my proper server(s). But they're not doing anything interesting, it was more a case of having used WAMP previously so I went with Apache when I moved my server to Ubuntu.

I only discovered how to use Proxies with Apache quite recently, really great idea in general. Passing specific subdomain requests to other servers seamlessly.

Apache gets the job done, and it's pretty much on all our client's low-cost shared web hosts. I never recommend that clients change their hosting, unless they are having major issues or our proposed solution can't work on it, then it's usually just to switch from IIS to Linux +apache (because those are the choices).

Maybe in future I'll be in a position to provision a more performant solution, where that's a requirement, and I'll consider the alternatives more carefully. I did install lighttpd on a virtual server recently and it was easier to bring up, and (seemingly) speedier than apache. But it doesn't appear to be available on any shared providers, so...

Thing is, since apache really is so widely used then going 'off the reservation', so to speak, could create problems for anyone who eventually inherits the server admin. So there's that.

Other thing I'm pondering is where containerization is going to take this whole business. I took a spin at it a few months ago, and it appears to be more geared to large-scale hosting providers.

Nginx has been available in Debian since 5.0 (I think) and Ubuntu since at least 12.04 and I'm sure the same is true of Redhat and other distros, so it's hardly going off reservation anymore. Installing and upgrading it is now no more difficult than installing Apache.

As long as you have free reign to install packages I would totally encourage you try it out and see what it can offer.

I might do that out of idle curiosity, one fine day.

 :-{)
 
Virtualbox. Hello?

Yes, and now the wonderful flawless nginx is around EVERYONE should instantly ditch Apache, all their years of experience, scripts, and configuration, and jump across to nginx which can be learnt instantly, provides all the same functionality, is perfectly compatible in every way, and will make scented rainbows shoot out their arse.

Or perhaps you could stop acting like a twattish fanboy and accept that they each have different benefits and drawbacks, and - even if nginx was objectively better in every way - it still takes time and effort to switch such a central piece of software, and so blurting about people "still" using what is by far the most used web server is simply pitiful.

I didn't say nginx was flawless and I didn't say everyone should switch.

Experience can be re-learnt, scripts re-written or better yet, written to be independent of the web-server, software and as for configuration, that should be handled by Saltstack or Puppet so you can test, re-test and deploy configuration changes safely. I know fully well that switching out takes time and effort, but that doesn't mean you shouldn't do it.

Also, I don't think nginx ships with the rainbow arse shooting module compiled in. Apache definitely ships with the false sense of superiority due to having more users module though :P

You two are funny

> I didn't say everyone should switch.

So what exactly is "People still use Apache?" saying, because it looks an awful lot like an expression of surprise at there being a plural of person not having switched away from Apache yet.

> ...[waffle]...

Missing the point.

> switching out takes time and effort, but that doesn't mean you shouldn't do it.

That's precisely what it means: When one has a long list of tasks to achieve, the solution is basically to prioritise those tasks in terms of how much benefit they provide versus how much it costs to achieve them.

Those that wont benefit from switching or for whom it occupies too much time should not do it, unless either/both of those cases change.

Whilst there are sacrifices nginx makes in favour of performance, there will remain people who require functionality it does not provide, and Apache is likely to remain their best/only option for quite a while. Assuming they aren't in the ~15-20% that (OMG) don't use nginx or Apache.

If you're still reading anything here as superiority then I'm probably wasting my time, but at least it's making Truffles happy.

Its called a generalisation, with an emoticon to show mock surprise.

You can't simply cut out bits of my post and claim it's waffle. You need to show which bit is waffle and why.

Also, why does something taking time and effort mean you shouldn't do it? When you have a laundry list of tasks that you need to prioritise, investigating change should be in there. Not necessarily at the top, but to negate it entirely is foolish.

There's only one cure for nerd rage: disco.
http://youtu.be/u1qN6gLbUMw

> You can't simply cut out bits of my post and claim it's waffle

I think you'll find I can - see msg:41579.25 for an example.

> Also, ...[waffle]... to negate it entirely is foolish.

It seems I mis-judged when reducing; my original wording made it impossible to claim inferrance of that nonsense. Oh well.

Keeping abreast of the available options is a relatively simple task that all competent professionals do, and an entirely separate step from deciding whether any alternative should be even considered, nevermind actually testing it out and scheduling when a live implementation is viable and sensible.

> [some waffle Pete said that I'm not going to quote properly so people don't know what I'm referencing to try and make me look better, but really all I'm doing is showing I can't be bothered to argue his points properly].

Pizza, what do you have on it?

Its Sunday night fever, but close enough.

Wow, who pissed on your waffles this morning?

I know far less about this stuff than you or Matt and my use of dedicated web servers these days is pretty much limited to reverse proxying node apps.

But Apache was always a struggle for me. Its config files felt arcane and I never really understood what I was doing with it.

Nginx feels like a breeze in comparison. I understand its config and its far easier to get it to do the (limited, admittedly) things I want it to do.

From my less technical perspective (though having used both a reasonable amount) my impression is that Nginx can do about 90% of what Apache can do and it does much of it better.

Now that I'm living in a country that uses the "ng" sound in a lot of the place names, "nginx" is a really hard word to read.

Yes, and ...? ¬_¬

The sole point I'm trying to get through the apparently lead-lined skulls in this place is that the remark "People still use X? Oh." is at best lame and non-constructive fanboy trolling. If you consider that X has more than twice the users of its nearest alternative, and will likely continue to have the largest user base of its class for the next five years, it makes the comment an incredibly dumb one too.

Whether or not X is or was any good is entirely irrelevant to that.

Why the fuck is it so hard to get people using their brains?

Hyperbole and facetiousness.

(i.e. it was less a statement of what is the case and more a statement of what Matt believes should be the case. That most people in most cases would probably be better served by Nginx than Apache. And he's probably right).