Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

StackOverflow Update: 560M Pageviews a Month, 25 Servers

highscalability.com

Post reply on HN
Print view View original

231–240 of 278 posts

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#231

Post by kasey_junk »

mseebach wrote: post #168
tomblomfield wrote: post #4

"One problem is not many tests. Tests aren’t needed because there’s a great community... If users find any problems with it they report the bugs that they’ve found." I'm often surprised at the paucity of test-coverage in relatively large companies. http://nerds.airbnb.com/testing-at-airbnb/

I think we're going trough a thesis/anti-thesis cycle on tests - in the beginning, there was militant testing, 100% coverage, testing getters and setters etc (as well as more complex stuff, obviously). Then some people started coming around to the idea that there are actually large swathes of code that is simple enough that testing doesn't actually add much value especially compared to the effort of writing them, the…

Actually, in the beginning there was no testing and even justifying having automated test infrastructure or spending developer resources on unit tests was a hard sell to management (what are they paying that QA dept. for after all).

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#232

Post by raverbashing »

ZoFreX wrote: post #218
noelwelsh wrote: post #207

StackOverflow shows just how powerful using a fast language can be. Compare the TechEmpower benchmarks for, say, Java vs Rails on a very simple JSON serialization benchmark (we can assume .Net would achieve comparable performance to the JVM): http://www.techempower.com/benchmarks/ The Java servers get around 900K requests/s on beefy hardware while Rails squeezes out 6K. That's a 150x difference! Any real application…

> That's a 150x difference! Any real application will be slower than this How many applications are rate limited by the speed of the front-end language? Not that many; the speed of the backing store will usually be the bottleneck.

Unless of course, the front-end processing takes so much time because it's complex and/or the language is slow.

One example, Django template processing is slow

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#233

Post by thomasz »

virmundi wrote: post #192
louthy wrote: post #135

Earlier quoted context omitted.

I assume it means using C# more as a functional language than an OO language, so that most garbage collection is of short-lifetime objects and therefore cheaper to collect (generation 0).

I take that statement to mean C# as C. Functional languages are terrible for garbage collection. Static methods are great since they are static, loaded once and probably in-lineable by the compiler/JIT.

The usual wisdom in clr-land is that the generational garbage collector can cope just fine with a lot of small, short lived objects. It absolutely doesn't like big, long lived objects that reach G1, G2 or the large object heap and must be traversed for a full gc.

This means that normally, you want to avoid data structures that introduce a lot of objects trees, linked lists and stuff like that. In some circumstances you might lose some speed, but win big time when the full GC comes. Additionally, It may be worthwhile to consider avoiding heap allocations by organizing your data as value types instead of ordinary objects.

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#234

Post by jacquesm »

shepik wrote: post #228
jacquesm wrote: post #143

Earlier quoted context omitted.

Programmer time is a factor that comes into play when you build something and re-training your people to use a stack they are not familiar with + the associated mistakes can get expensive in a hurry. If you happen to be familiar with the MS stack and are able to factor the license fees into your business plan there is no strong reason against and many for using that particular stack. MS programmers tend to be a bit m…

> MS stack is more performant than the Linux one on the same hardware Could you please provide some evidence to support that?

Without breaking any promises to others here is an article with conclusions quite similar to mine:

http://www.webperformance.com/load-testing/blog/2011/11/what...

Keep in mind that it hardly ever is the webserver that is the bottle neck.

My theory on why this is the case is very simple: MS can afford to throw vast amounts of money at optimizations that are next to impossible in Linux simply because the coupling between the layers in Linux is looser. And that's a good thing, it translates into better security and fewer bugs.

As always, optimization alone is not a reason enough to go down a certain path. But for raw speed on requests it's fairly hard to beat IIS, if that's what you're after (I'm usually not, and even when it matters I can comfortably saturate most outbound links from a single server doing light processing, and as soon as the processing becomes the bottle-neck the CPU cache size, RAM and so on matter more than your OS, but given identical hardware a 'dirty' approach should yield better results at a cost of unreliability/complexity).

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#235

Post by jaequery »

liotier wrote: post #227
dekhn wrote: post #162

Earlier quoted context omitted.

You can negotiate with AWS.

I don't believe that startups in their garage days have anywhere near the necessary leverage to negotiate anything with Amazon...

Agree, not many people actually realize just how slow and expensive AWS actually is. Kudos to SO for going with their own hardware.

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#236

Post by Nick-Craver »

thomasz wrote: post #233
virmundi wrote: post #192

Earlier quoted context omitted.

I take that statement to mean C# as C. Functional languages are terrible for garbage collection. Static methods are great since they are static, loaded once and probably in-lineable by the compiler/JIT.

The usual wisdom in clr-land is that the generational garbage collector can cope just fine with a lot of small, short lived objects. It absolutely doesn't like big, long lived objects that reach G1, G2 or the large object heap and must be traversed for a full gc. This means that normally, you want to avoid data structures that introduce a lot of objects trees, linked lists and stuff like that. In some circumstances y…

It depends on your definitions of "a lot" is. It can scale pretty high by default, but when you're doing hundreds of millions of objects in a short window, you can actually measure pauses in the app domain while GC runs. These have a pretty decent impact on requests that hit up against that block.

The vast majority of programmers need not worry about this, but if you're putting millions of objects through a single app domain in a short time, then it's a concern you should be aware of. This applies to pretty much any managed language.

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#237

Post by cagenut »

Nick-Craver wrote: post #155
MichaelGG wrote: post #55

So this is what, a 2000 request/sec peak? Over 11 servers, that's like 200 requests/sec peak per frontend? The problem with scale-up is if you actually have to get a few times larger, it becomes super expensive. But fortunately hardware is increasing so much that you can probably just get away with it now. There's probably a crossover point we're rapidly approaching where even global-scale sites can just do all their…

Peak is more like 2600-3000 requests/sec on most weekdays. Remember that programming, being a profession, means our weekdays are significantly busier than weekends (as you can see here: https://www.quantcast.com/p-c1rF4kxgLUzNc ). It's almost all over 9 servers, because 10 and 11 are only for meta.stackexchange.com, meta.stackoverflow.com, and the development tier. Those servers also run around 10-20% CPU which means…

  I actually just wrote a lengthy email about what we're planning for storage on one of our SQL clusters...perhaps I should echo it verbatim as a blog post? I'm not sure how many people care about that sort of stuff outside our teams.
Yes! please yes. The occasional real world gem like that is what makes scrolling past a thousand javascript-framework/obscure-language/alternative-datastore posts here worth it (for me).

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#238

Post by valarauca1 »

kondro wrote: post #185
jaequery wrote: post #178

Earlier quoted context omitted.

Same server in AWS, it'll cost $50k+/month for the same thing. SO just bought it one time fee for probably around $30k or so, I'll say they are doing something right!

The don't have servers with that much RAM, but a 244GB RAM server with 8x800GB SSDs and 32 cores costs <$5k a month ad-hoc and about $1300/month with a 3 year commitment.

$1300/month w/ a 3 year commitment is $46.8k over the lifespan. Which is more then 4x more then the actual hardware cost. Your ROI on a physical server with the same stats at that price tag would be 8 months.

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#239

Post by phkahler »

Shog9 wrote: post #184
phkahler wrote: post #91

I haven't been able to use SO since they shut off myopenid last year. I can't log in, so can't change my credentials ;-( Any ideas HN?

You can email us (as George said) or you can just punch your email address into the account recovery tool and get an automated email telling you what to click within the space of a few minutes: http://stackoverflow.com/users/account-recovery We... probably need to make that link a bit more obvious.

Thanks, that was easy. So now I just have a password instead of linking to another account @facebook or google+ I guess that works ;-)

Re: StackOverflow Update: 560M Pageviews a Month, 25 Servers

#240

Post by phkahler »

GeorgeBeech wrote: post #97
phkahler wrote: post #91

I haven't been able to use SO since they shut off myopenid last year. I can't log in, so can't change my credentials ;-( Any ideas HN?

Email team@stackexchange.com they should be able to help you out. (I'm an SE SysAdmin )

Thanks, I followed the recovery link provided by another SE person.
Post reply on HN