Hello,
I’ve been getting this error a lot in my testing. I never received it
with Apache. I understand what the message say but not why it’s coming
about. My headers being sent have length. Or is this message referring
to content returned by PHP at the server end? I’m sure one of you
experts here knows what’s going on with this. I only get it on POST
requests and not most of my ajax calls (which are GET).
Thanks for any help with tracking this down. I’m seeing it in Firebug
and when it occurs I cannot see the real response content - even though
content is indeed returned as well since it has effect on my javascript
code and updates the screen. Weird.
Chris 
Response from server:
411 Length Required
411 Length Required
nginx/0.6.32
Header sent to server using ajax (only removed some identity info, HOST
line):
User-Agent Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1)
Gecko/2008072820 Firefox/3.0.1
Accept text/javascript, text/html, application/xml, text/xml, /
Accept-Language en-us,en;q=0.5
Accept-Encoding gzip,deflate
Accept-Charset ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive 300
Connection keep-alive
X-Requested-With XMLHttpRequest
Content-Type application/x-www-form-urlencoded; charset=utf-8
Referer http://www.xxxxx.com/
Content-Length 269
Cookie ZZZZSession=3ce34bb3dbc257ff61405175f941f1d7
Pragma no-cache
Cache-Control no-cache
On Sun, Aug 17, 2008 at 02:14:18PM +0700, Chris S. wrote:
code and updates the screen. Weird.
Chris 
Set error_log level to info:
error_log /path/to/log info;
Then nginx will log a reason. Now there are 3 reasons:
“client sent invalid “Content-Length” header”
“client sent … method without “Content-Length” header”
“client sent “Transfer-Encoding: chunked” header”
The last reason is nginx feature: it still does not support chunked
body.
Thanks Igor.
I turned on info and it says the reason is client didn’t send
“Content-Length” header. Now, why would Firefox (3) not do that? Is this
common?
I noticed in Firebug that the header sent ALSO has “Transfer-Encoding:
chunked”.
I had the same problem with Lightpd and using a Flash uploader but they
refuse to fix it to ignore chunked, so I could not use it with Flash.
One reason I switched to using nginx is that unlike Lighttpd it works
with the flash uploader.
So should I just ignore this as it still works as expected? But I
cannot see content in Firebug, or maybe I need to force a header in JS
code (maybe the mootools JS library mistakenly doesn’t set the Content
Length hmm, I should explore that too.) Just a minefield. How is it that
the server returns this error page but also I am getting the JSON data
back as well?
Chris
My mistake on that last message. I read the wrong headers info.
It seems ngninx reports no Content-Length header was sent but according
to Firebug these are the headers - note it says Content Length is
present and chunked not mentioned…
Host www.xxxx.com
User-Agent Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1)
Gecko/2008072820 Firefox/3.0.1
Accept text/javascript, text/html, application/xml, text/xml, /
Accept-Language en-us,en;q=0.5
Accept-Encoding gzip,deflate
Accept-Charset ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive 300
Connection keep-alive
X-Requested-With XMLHttpRequest
Content-Type application/x-www-form-urlencoded; charset=utf-8
Referer http://www.xxxx.com/
Content-Length 120
Cookie XXXXXSession=3ce34bb3dbc257ff61405175f941f1d7
Pragma no-cache
Cache-Control no-cache
On Sun, Aug 17, 2008 at 08:53:55PM +0700, Chris S. wrote:
Accept-Encoding gzip,deflate
Accept-Charset ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive 300
Connection keep-alive
X-Requested-With XMLHttpRequest
Content-Type application/x-www-form-urlencoded; charset=utf-8
Referer http://www.xxxx.com/
Content-Length 120
Cookie XXXXXSession=3ce34bb3dbc257ff61405175f941f1d7
Pragma no-cache
Cache-Control no-cache
Then only nginx debug log or tcpdump will help.
Ok. I used Wireshark and tracked this down. I’m posting here so that
others will know why this happens. It is a result of poor behaviour in
Firebug.
In fact, if you click on the log tab to view request data and headers
then Firebug actually re-submits the POST data a second (hidden) time to
provide data to show you. Of course, since POST methods will update
server data and could insert new database rows this is very bad of
Firebug. It is treating the POST like a GET in this case where it should
have logged the original POST reply data instead. I have posted a bug in
their google group and hopefully someday they will give it some
attention.
When Firebug submits the second time it does NOT include the
Content-Length header as it did at first, but it shows as though it did.
This is why my page was correctly updated but an 411 error was shown in
the Firebug log. The original response was not saved and the error was
from the repeated POST request.
Just something for people to be aware of.
Chris
On Sun, Aug 17, 2008 at 08:40:39PM +0700, Chris S. wrote:
One reason I switched to using nginx is that unlike Lighttpd it works
with the flash uploader.
So should I just ignore this as it still works as expected? But I
cannot see content in Firebug, or maybe I need to force a header in JS
code (maybe the mootools JS library mistakenly doesn’t set the Content
Length hmm, I should explore that too.) Just a minefield. How is it that
the server returns this error page but also I am getting the JSON data
back as well?
nginx tests Content-Length absence before Transfer-Encoding, thus
it logs this error, although the request is valid: it has either
Transfer-Encoding or Content-Length.
On Fri, Nov 28, 2008 at 09:46:34AM +0100, Rocky Zhang wrote:
Here is a solution:
KingTch.com is for sale | HugeDomains
The patch is wrong, it simply drops chunked body.
Hi all
I’m having a similar issue. In Firebug, the Request Headers
Content-Length shows 112. However, I got the 411 Content Length Required
response.
Does any of the experts here know what’s going on? Thanks in advance!
Request Headers
Host www.xxxxx.com
User-Agent Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.5; en-US;
rv:1.9.0.1) Gecko/2008070206 Firefox/3.0.1
Accept text/html,application/xhtml+xml,application/xml;q=0.9,/;q=0.8
Accept-Language en-us,en;q=0.5
Accept-Encoding gzip,deflate
Accept-Charset ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive 300
Connection keep-alive
X-Requested-With XMLHttpRequest
Content-Type application/x-www-form-urlencoded; charset=UTF-8
Referer xxxxx.com
Content-Length 112
Cookie
_xxxxxx_session=BAh7CjoPbGFzdF9sb2dpbnU6CVRpbWUNhh0bgAAAEI86EHJlcXVlc3RfdXJp%250AMDoMY3NyZl9pZCIlNWM1ZmYxZjlmZWQ2YmJmNTM2YWJlOGQwOGE0MjhkZDIi%250ACmZsYXNoSUM6J0FjdGlvbkNvbnRyb2xsZXI6OkZsYXNoOjpGbGFzaEhhc2h7%250AAAY6CkB1c2VkewA6CXVzZXJvOjVBY3RpdmVSZWNvcmQ6OkFzc29jaWF0aW9u%250Aczo6QmVsb25nc1RvQXNzb2NpYXRpb24JOgtAb3duZXJvOgtEZXRhaWwIOgxA%250AcGVyc29uQAs6EEBhdHRyaWJ1dGVzewsiFXdvcmxkX3Zpc2liaWxpdHkiD2V2%250AZXJ5dGhpbmciFWluZm9ybWF0aW9uX3R5cGUiEUVtYWlsQWRkcmVzcyIPdXBk%250AYXRlZF9vbiIYMjAwOC0wOC0wNCAwNTo0OToxNSIHaWQiBjIiE2luZm9ybWF0%250AaW9uX2lkIgYxIg5wZXJzb25faWQiBjE6FkBhdHRyaWJ1dGVzX2NhY2hlewA6%250ADEBsb2FkZWRUOgxAdGFyZ2V0bzoJVXNlchc6JUBkZXRhaWxfbWVtYmVyc2hp%250AcF9yZWxhdGlvbnNoaXBzMDopQHVzZXJfcHJvZmlsZV9lbWFpbF9hZGRyZXNz%250AX3JlcXVlc3RzMDoTQGhpc3RvcnlfaXRlbXMwOxB7CiIPYWNjb3VudF9pZCIG%250AMSIVb3B0aW9uYWxfZGF0ZV9pZDAiCXR5cGUiCVVzZXIiB2lkIgYxIgxuYW1l%250AX2lkIgYxOg1AYWNjb3VudDA6KEB1c2VyX3Byb2ZpbGVfZW1haWxfYWRkcmVz%250Ac19yZXF1ZXN0MDoNQGRldGFpbHMwOiNAZGV0YWlsX2VtYWlsX2FkZHJlc3Nf%250AcmVxdWVzdHMwOhBAYmlydGhfZGF0ZTA6FEBncm91cHNfbWFuYWdlZDA6HEBl%250AbWFpbF9hZGRyZXNzX3JlcXVlc3RzMDodQG1hbmFnZXJfc3RhdHVzX3JlcXVl%250Ac3RzMDsRewA6DEBncm91cHMwOitAZGVsZXRlX3ByaW1hcnlfZW1haWxfYWRk%250AcmVzc19yZXF1ZXN0czA6EUBtZW1iZXJzaGlwczA6IUBncm91cF9tYW5hZ2Vy%250AX3JlbGF0aW9uc2hpcHMwOgpAbmFtZTA6EEByZWZsZWN0aW9ubzo0QWN0aXZl%250AUmVjb3JkOjpSZWZsZWN0aW9uOjpBc3NvY2lhdGlvblJlZmxlY3Rpb24NOhhA%250AdGhyb3VnaF9yZWZsZWN0aW9uRjoTQGFjdGl2ZV9yZWNvcmRjC0RldGFpbDoQ%250AQGNsYXNzX25hbWUiC1BlcnNvbjoLQGtsYXNzYwtQZXJzb246C0BtYWNybzoP%250AYmVsb25nc190bzoNQG9wdGlvbnN7ADskOgtwZXJzb246FkBwcmltYXJ5X2tl%250AeV9uYW1lIg5wZXJzb25faWQ%253D–2bd2eb7eff4dcf6d5f1b8290df567298516fbfe9;
connection=efa9ce4469ee3f84e8488d939c6e82d4; email_address=xxx%40xxx.com
Chris S. wrote:
Ok. I used Wireshark and tracked this down. I’m posting here so that
others will know why this happens. It is a result of poor behaviour in
Firebug.
In fact, if you click on the log tab to view request data and headers
then Firebug actually re-submits the POST data a second (hidden) time to
provide data to show you. Of course, since POST methods will update
server data and could insert new database rows this is very bad of
Firebug. It is treating the POST like a GET in this case where it should
have logged the original POST reply data instead. I have posted a bug in
their google group and hopefully someday they will give it some
attention.
When Firebug submits the second time it does NOT include the
Content-Length header as it did at first, but it shows as though it did.
This is why my page was correctly updated but an 411 error was shown in
the Firebug log. The original response was not saved and the error was
from the repeated POST request.
Just something for people to be aware of.
Chris
I’ve updated the patch in this way. you can check it in the same blog
post.
Rocky Zhang wrote:
Yes. If there is no chunked body, it works fine. If there is, we could
check its request header information and close it if we haven’t
implemented chunked body feature. Could we patch it?
Igor S. wrote:
On Fri, Nov 28, 2008 at 09:46:34AM +0100, Rocky Zhang wrote:
Here is a solution:
KingTch.com is for sale | HugeDomains
The patch is wrong, it simply drops chunked body.
On Fri, Nov 28, 2008 at 04:26:29PM +0100, Rocky Zhang wrote:
Yes. If there is no chunked body, it works fine. If there is, we could
check its request header information and close it if we haven’t
implemented chunked body feature. Could we patch it?
I do not understand what yuo want to get.
Currently nginx does not support a chunked request body and returns
411 error code for such requests.
Also nginx returns the 411 error if a POST/PUT request has no
“Content-Length”.
If the request has no body, then it must send “Content-Length: 0”.
Unfortunately I found that when I use POST method to request an AJAX
function, the browser sends a request with neither body nor
“Content-Length: 0”.
Igor S. wrote:
On Fri, Nov 28, 2008 at 04:26:29PM +0100, Rocky Zhang wrote:
Yes. If there is no chunked body, it works fine. If there is, we could
check its request header information and close it if we haven’t
implemented chunked body feature. Could we patch it?
I do not understand what yuo want to get.
Currently nginx does not support a chunked request body and returns
411 error code for such requests.
Also nginx returns the 411 error if a POST/PUT request has no
“Content-Length”.
If the request has no body, then it must send “Content-Length: 0”.
It sounds like that browser is not spec compliant. You may have to
fake a body by sending a form encoded parameter just to generate a
dummy body.
Dave
On Tue, Dec 02, 2008 at 01:34:36AM +1100, Dave C. wrote:
It sounds like that browser is not spec compliant. You may have to
fake a body by sending a form encoded parameter just to generate a
dummy body.
It seems that Apache allowed always such requests (POSTs without
Content-Length). I’m not sure should I add this ?
I think AJAX will work well with nginx if you add this. Maybe my patch
may do a little thing for you? FYI, this problem will happen on both IE
7 and Firefox 3.0.4.
Igor S. wrote:
On Tue, Dec 02, 2008 at 01:34:36AM +1100, Dave C. wrote:
It sounds like that browser is not spec compliant. You may have to
fake a body by sending a form encoded parameter just to generate a
dummy body.
It seems that Apache allowed always such requests (POSTs without
Content-Length). I’m not sure should I add this ?
Yes. If there is no chunked body, it works fine. If there is, we could
check its request header information and close it if we haven’t
implemented chunked body feature. Could we patch it?
Igor S. wrote:
On Fri, Nov 28, 2008 at 09:46:34AM +0100, Rocky Zhang wrote:
Here is a solution:
KingTch.com is for sale | HugeDomains
The patch is wrong, it simply drops chunked body.
It seems that Apache allowed always such requests (POSTs without
Content-Length). I’m not sure should I add this ?
Looking into the RFC’s it looks like Apache might be doing the right
thing.
RFC1945:
A valid Content-Length is required on all HTTP/1.0 POST requests. An
HTTP/1.0 server should respond with a 400 (bad request) message if it
cannot determine the length of the request message’s content.
RFC2616:
The presence of a message-body in a request is signaled by the
inclusion
of a Content-Length or Transfer-Encoding header field in the
request’s
message-headers
So,
POST /someurl HTTP/1.1
Host: whatever
\r\n
\r\n
Is a valid post request, but only in the case that the request does not
contain a body. If the post does contain a body, then it must be
identified
as such with a Content-Lenth: header.
Cheers
Dave
On Mon, Dec 1, 2008 at 6:38 AM, Igor S. [email protected] wrote:
It seems that Apache allowed always such requests (POSTs without
Content-Length). I’m not sure should I add this ?
I can vouch that we’ve been bitten by that before with AJAX, during
switchovers from straight Apache to nginx reverse proxying for Apache.
We’ve got a much bigger switchover in the works and I’ve though about
removing NGX_HTTP_POST from src/http/ngx_http_request.c (line 1414 in
0.7.22, so it’d just apply to PUT). Having a switch to turn off the No
Content Length check for POSTs would be most excellent.
On Mon, Dec 01, 2008 at 05:08:50PM -0800, Mark M. wrote:
Content Length check for POSTs would be most excellent.
I thought about the same, here is patch.