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

The author of Nginx on why V8 is not suitable for web servers

translate.google.com

Post reply on HN
Print view View original

61–70 of 75 posts

Re: The author of Nginx on why V8 is not suitable for web servers

#61

Post by otterley »

JoachimSchipper wrote: post #53
ErikCorry wrote: post #45

Speaking only for myself as ever: Handling out of memory errors is a hard problem. At the point when you can't allocate more memory you are busted in most languages. For example in C a slightly deeper function call depth can cause the stack to expand at any point. If that point coincides with you running out of memory then you are busted. No program can keep running with a broken stack. If V8 were capable of recoveri…

You're being too clever - if you run out of memory, throw away the request and all associated resources and return "500: Internal server error". There are lots of reasons why this isn't perfect (if the OS overcommits memory, like Linux (typically?) does, it will have to kill you to avoid running out; if you need to use swap you're too slow to be useful; etc), but it's not impossible. node.js (one process for many req…

From a service perspective, I'd prefer to have the process crash and supervisor restart it. Returning HTTP/500 errors on memory allocation failures (which usually last until the process is restarted anyway) is likely to deny service to a larger number of users.

Re: The author of Nginx on why V8 is not suitable for web servers

#62

Post by wazoox »

baguasquirrel wrote: post #2

It's scary what you can do with machine translation these days...

Indeed. However though Google translate quite correctly from most common languages to English, translation to other languages (even English to French or English to Spanish) often hardly makes any sense.

Re: The author of Nginx on why V8 is not suitable for web servers

#63

Post by delinka »

I keep seeing these replies on how well Google Translate performed. I didn't realize while reading that it was machine-translated, but my immediate thought was "this person is not a native English speaker or needs much work his attempt at Yoda-speak does."

Re: The author of Nginx on why V8 is not suitable for web servers

#64

Post by seats »

delinka wrote: post #63

I keep seeing these replies on how well Google Translate performed. I didn't realize while reading that it was machine-translated, but my immediate thought was "this person is not a native English speaker or needs much work his attempt at Yoda-speak does."

But still understandable, so that's a success I'd say.

I remember when nginx first came out (the web server that the author built) and the biggest barrier to adoption was the fact that all the documentation was in russian. Online machine translation did not produce understandable content at that time.

Re: The author of Nginx on why V8 is not suitable for web servers

#65

Post by merloen »

Sephr wrote: post #20
naner wrote: post #18

Earlier quoted context omitted.

It likely was not completely translated by machine. Google Translate lets anyone reading the translation edit it to fix errors. Highlight some text and notice the link that appears allowing you to "Contribute a better translation" for that particular passage.

It's more likely that it only helps train Google Translate, and doesn't actually apply any changes directly to the translated text. Otherwise vandalism would be way too easy.

Why on earth is the comment by Pooter dead?

Re: The author of Nginx on why V8 is not suitable for web servers

#66

Post by aaronblohowiak »

ErikCorry wrote: post #45

Speaking only for myself as ever: Handling out of memory errors is a hard problem. At the point when you can't allocate more memory you are busted in most languages. For example in C a slightly deeper function call depth can cause the stack to expand at any point. If that point coincides with you running out of memory then you are busted. No program can keep running with a broken stack. If V8 were capable of recoveri…

If you are using GBs of Heap space, GC pauses are going to make you cry before you worry about an OOM error. More, smaller processes are the way to go for a number of reasons (recovery is cheaper when you lose smaller things..)

Re: The author of Nginx on why V8 is not suitable for web servers

#67

Post by daleharvey »

ballard wrote: post #27

Node.js is often used because it is cool, new and easier for front-end developers to develop a functioning backend. For more traditional uses that want something less hackish, erlang for example, crashes only a single thread not the entire vm. Functional programming languages in general like haskell and erlang are interesting for backend core services.

memory allocation failures crash the erlang vm

Re: The author of Nginx on why V8 is not suitable for web servers

#68

Post by EGreg »

EGreg wrote: post #26

Sysoev brings up a couple good points that affect Node.js . 1) The fact memory allocation might crash the process is a serious problem. Is this still the case? 2) The fact that garbage collection is still stop-the-world may be a problem for server availability. Are we able to call the generational collector from node, or no? Things that don't involve node: The ability to create multiple objects in different contexts.…

upvotes are great :) But does anyone know the answers to 1) and 2)?

Re: The author of Nginx on why V8 is not suitable for web servers

#69

Post by _cbsz »

JoachimSchipper wrote: post #53
ErikCorry wrote: post #45

Speaking only for myself as ever: Handling out of memory errors is a hard problem. At the point when you can't allocate more memory you are busted in most languages. For example in C a slightly deeper function call depth can cause the stack to expand at any point. If that point coincides with you running out of memory then you are busted. No program can keep running with a broken stack. If V8 were capable of recoveri…

You're being too clever - if you run out of memory, throw away the request and all associated resources and return "500: Internal server error". There are lots of reasons why this isn't perfect (if the OS overcommits memory, like Linux (typically?) does, it will have to kill you to avoid running out; if you need to use swap you're too slow to be useful; etc), but it's not impossible. node.js (one process for many req…

Overcommit seems to have changed somewhat since the "early days," as it's now heuristic. I ran into this once when I couldn't malloc 8 GB on a 64-bit machine because RAM + swap was less than 8 GB. http://www.mjmwired.net/kernel/Documentation/vm/overcommit-a...

Re: The author of Nginx on why V8 is not suitable for web servers

#70

Post by koenigdavidmj »

imurray wrote: post #47
JulianMorrison wrote: post #44

IIRC, on Linux at least the memory allocation thing is moot. If your process runs out of memory, it dies anyway. You can't catch "malloc says no" because malloc never says no. It either says yes or blows your process's brains out.

"malloc never says no" malloc says no if you set a ulimit in your shell. I do this on the desktop to stop myself from shooting myself in the foot. (Recent kernels also let you turn off the overcommitting behavior.)

Note to self: do not leave root terminal behind xlock, since I have no guarantee that xlock will stay around.
Post reply on HN