Using Fapws to handle many concurent big uploads

[sys]

I am thinking about Fapws to serve uploads (200Mb +) for some
highloaded projects.
The problem now is that nginx or lighttpd buffers whole user request
and then sends it to back and. This couse additional disk io
operations. I just want to catch uploaded file stream, parse it and
save to storage.

The question is:
Is it possible to serve uploads without worker processes locks?

[Dipankar Sarkar]

You can try using the NGINX upload module to do what you just said. It will do all the parsing and save it to storage.

http://brainspl.at/articles/2008/07/20/nginx-upload-module

I think its a better solution that using fapws

regards,
Dipankar Sarkar

email: m...@dipankar.name | web: http://dipankar.name | blog: http://desinerd.com
Mobile (India) +91 981 145 2098 | Phone (US) +1 415 670 9162
twitter : http://twitter.com/dipankarsarkar | skype : dipankar.sarkar

[william]

Fapws use libev events, thus if you are not using blocking processes
you will not be locked.

Easiest way to not lock the webserver is to use files. Fapws manage
files as chunks of data. Each chunks is managed separatly. That way
you can process lot of big files.

The size of those chunks is defined by the variable: MAX_BUFF

William

[Kostya Pereverzev]

I found that whole request is collected in memmory before passing it
to WSGI Environment.