I'm running Cherokee Web Server 1.0.9

I've give it a try with uWSGI 0.9.7-dev

Got a 504 Gateway Timeout...


Here is my actual ini file :
[uwsgi]
socket=127.0.0.1:32944
home=/home/administrator/.virtualenvs/django.test.0.0.4
pythonpath=/home/administrator/sandbox/demo/projects/
pythonpath=/home/administrator/sandbox/demo/projects/demo/
env = DJANGO_SETTINGS_MODULE=demo.staging_settings
module = django.core.handlers.wsgi:WSGIHandler()
processes=6
master=true
harakiri=false
disable-logging=true
limit-as=128
listen=128
socket-timeout=240
memory-report=true
logto=/home/administrator/sandbox/demo/log/uwsgi.log

Here is the log :
*** Starting uWSGI 0.9.7-dev (32bit) on [Wed Dec  8 19:01:50 2010] ***
compiled with version: 4.4.3
limiting address space of processes...
your process address space limit is 134217728 bytes (128 MB)
your memory page size is 4096 bytes
Python version: 2.6.5 (r265:79063, Apr 16 2010, 13:28:26)
[GCC 4.4.3]
Setting PythonHome to /home/administrator/.virtualenvs/django.test.0.0.4...
added /home/administrator/sandbox/demo/projects/ to pythonpath.
added /home/administrator/sandbox/demo/projects/demo/ to pythonpath.
uwsgi socket 0 inherited INET address 127.0.0.1:32944 fd 3
your server socket listen backlog is limited to 128 connections
*** Operational MODE: preforking ***
WSGI application 0 (SCRIPT_NAME=) ready on interpreter 0x82abcb0 (default
app)
*** uWSGI is running in multiple interpreter mode ***
gracefully (RE)spawned uWSGI master process (pid: 674)
spawned uWSGI worker 1 (pid: 3787, cores: 1)
spawned uWSGI worker 2 (pid: 3788, cores: 1)
spawned uWSGI worker 3 (pid: 3789, cores: 1)
spawned uWSGI worker 4 (pid: 3790, cores: 1)
spawned uWSGI worker 5 (pid: 3791, cores: 1)
spawned uWSGI worker 6 (pid: 3792, cores: 1)
SIGPIPE: writing to a closed pipe/socket/fd (probably the client
disconnected) on request /fr/admin/photologue/galleryupload/add/ !!!
write(): Broken pipe [plugins/python/wsgi_subhandler.c line 173]



2010/12/8 Roberto De Ioris <[email protected]>

> Hi roberto and thanks for your response.
> Unfortunally i'm still encountering problems with that, so I can't figure
> out what can be the perfect setup for my case.
> Putting post-buffering 4096 just don't work for me. It never upload the
> file
> and raise a browser error telling me that the connection with the server
> has
> been reinitialized.
> I've tried disabling harakiri with harakiri=false and it works fine unless
> I
> upload +/- 20 mb zip files...
>
> So I'm asking me what's the problem with post-buffering 4096. My virtual
> server runs with only 256mb can it be a clue ?
> Or is there something I can do from the cherokee setup part ?
> Also is disabling harakiri can leads into other problems in my app, what
> others of you use to setup a simple cms/blog app with big file uploads
> need
> (I need +/- 200 mb upload for video files) ?
>
> In place of passing arguments from command line i've done some ini files,
> to
> make the tests.
>
> Here is my last ini config file :
>
> [uwsgi]
> socket=127.0.0.1:32944
> home=/home/administrator/.virtualenvs/django.test.0.0.4
> pythonpath=/home/administrator/sandbox/demo/projects/
> pythonpath=/home/administrator/sandbox/demo/projects/demo/
> env = DJANGO_SETTINGS_MODULE=demo.staging_settings
> module = django.core.handlers.wsgi:WSGIHandler()
> processes=4
> master=true
> harakiri=false
> limit-as=128
> socket-timeout=360
> memory-report=true
> logto=/home/administrator/sandbox/demo/log/uwsgi.log
>
> And last log :
> {address space usage: 72519680 bytes/69MB} {rss usage: 56455168
> bytes/53MB}
> [pid: 1915|app: 0|req: 13/48] 192.168.0.13 () {62 vars in 1594 bytes} [Wed
> Dec  8 16:21:51 2010] GET /fr/admin/photologue/galleryupload/add/ =>
> generated 14100 bytes in 436 msecs (HTTP/1.1 200) 10 headers in 511 bytes
> (0
> async switches on async core 0)
> {address space usage: 40144896 bytes/38MB} {rss usage: 23916544
> bytes/22MB}
> [pid: 1917|app: 0|req: 12/49] 192.168.0.13 () {60 vars in 1410 bytes} [Wed
> Dec  8 16:21:52 2010] GET /admin/jsi18n/ => generated 4201 bytes in 23
> msecs
> (HTTP/1.1 200) 4 headers in 144 bytes (0 async switches on async core 0)
> SIGPIPE: writing to a closed pipe/socket/fd (probably the client
> disconnected) on request /fr/admin/photologue/galleryupload/add/ !!!
> write(): Broken pipe [pyutils.c line 101]
> {address space usage: 41504768 bytes/39MB} {rss usage: 25276416
> bytes/24MB}
> [pid: 1918|app: 0|req: 13/50] 192.168.0.13 () {62 vars in 1680 bytes} [Wed
> Dec  8 16:22:22 2010] POST /fr/admin/photologue/galleryupload/add/ =>
> generated 0 bytes in 23933 msecs (HTTP/1.1 302) 11 headers in 709 bytes (0
> async switches on async core 0)

It looks like cherokee is disconnecting after the upload is complete.
What version of cherokee are you using ?

Can you try with the current mercurial code as its post-buffering system
is a lot better ? (and if it isa cheroke problem it will log it)

--
Roberto De Ioris
http://unbit.it
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi
_______________________________________________
uWSGI mailing list
[email protected]
http://lists.unbit.it/cgi-bin/mailman/listinfo/uwsgi

Reply via email to