In our case, I've tracked the problem to the client library we are using,
ASIHTTPRequest. The library is occasionally sending an incorrect
Content-Length when configured to use persistent connections.
On Aug 10, 2010, at 8:47 AM, Roger wrote:
> I'm seeing the same symptoms. The rate is much
I'll try running under Cherokee + uWSGI and see if the problem shows
up again. Something tells me this is related to FCGI...
It's still hard to separate the false positives though. If a user
stops a upload, you see the exact same traceback.
On Aug 10, 12:47 pm, Roger wrote:
> I'm seeing the same
I'm seeing the same symptoms. The rate is much lower than 1 in 5 -
maybe 1 in 100 - but definitely the same error. Were you guys able to
make any progress on a solution?
On Aug 4, 1:17 am, hcarvalhoalves wrote:
> Now that someone else mentioned, yes, I believe we have the same
> problem.
>
> I
Now that someone else mentioned, yes, I believe we have the same
problem.
I run Django thru FastCGI to a Cherokee Web Server, and occasionally,
uploads fail to continue (the traceback shows that the code hanged at
consuming the input stream, then the connection got reset by the
client's browser, r
We have an intermittent problem when uploading files. About one in five
uploads fails, when the MultiPartParser receives an HTTP_CONTENT_LENGTH of zero.
We are running Django with lighttpd and fastcgi, has anyone else encountered
this problem?
--
Eric Chamberlain
--
You received this mes
5 matches
Mail list logo
