Commit Graph

4226 Commits

Author SHA1 Message Date
Igor Sysoev
fc35bf6434 release-1.1.5 tag 2011-10-05 14:46:09 +00:00
Igor Sysoev
6ef5a8d298 nginx-1.1.5-RELEASE 2011-10-05 14:44:11 +00:00
Igor Sysoev
290ad9c012 Adding RELEASE procedure in GNUmakefile. 2011-10-05 13:24:32 +00:00
Igor Sysoev
ab0078a4c8 Version bump. 2011-10-05 13:13:25 +00:00
Maxim Dounin
20139ff13a Fixed cache bypass caching of non-cacheable replies (ticket #21).
If cache was bypassed with proxy_cache_bypass, cache-controlling headers
(Cache-Control, Expires) wasn't considered and response was cached even
if it was actually non-cacheable.

Patch by John Ferlito.
2011-10-05 10:14:21 +00:00
Maxim Dounin
3aba768855 Added uwsgi_buffering and scgi_buffering directives.
Patch by Peter Smit.
2011-09-30 11:53:27 +00:00
Igor Sysoev
f3ae6a6102 Using strtod() instead of atofp() to support a lot of digits after dot in
"start" parameter value.
2011-09-30 09:18:28 +00:00
Igor Sysoev
a40e7eed30 Fix of building on platforms with 32-bit off_t. (closed #23) 2011-09-29 15:19:36 +00:00
Maxim Dounin
5cc90f309d Fixed segmentation fault with empty config on Windows.
See here for report:
http://mailman.nginx.org/pipermail/nginx-ru/2011-September/043288.html
2011-09-27 12:07:09 +00:00
Maxim Dounin
554768dabb Fix for "ssl_session_cache builtin" (broken since 1.1.1, r3993). 2011-09-27 12:06:07 +00:00
Maxim Dounin
886c6295ee Better handling of late upstream creation.
Configuration with duplicate upstream blocks defined after first use, i.e.
like

    server {
        ...
        location / {
            proxy_pass http://backend;
        }
    }

    upstream backend { ... }
    upstream backend { ... }

now correctly results in "duplicate upstream" error.

Additionally, upstream blocks defined after first use now handle various
server directive parameters ("weight", "max_fails", etc.).  Previously
configuration like

    server {
        ...
        location / {
            proxy_pass http://backend;
        }
    }

    upstream backend {
        server 127.0.0.1 max_fails=5;
    }

incorrectly resulted in "invalid parameter "max_fails=5"" error.
2011-09-27 11:18:51 +00:00
Maxim Dounin
b16918ed08 Cache: fix for sending of stale responses.
For normal cached responses ngx_http_cache_send() sends last buffer and then
request finalized via ngx_http_finalize_request() call, i.e. everything is
ok.

But for stale responses (i.e. when upstream died, but we have something in
cache) the same ngx_http_cache_send() sends last buffer, but then in
ngx_http_upstream_finalize_request() another last buffer is send.  This
causes duplicate final chunk to appear if chunked encoding is used (and
resulting problems with keepalive connections and so on).

Fix this by not sending in ngx_http_upstream_finalize_request()
another last buffer if we know response was from cache.
2011-09-27 11:17:11 +00:00
Maxim Dounin
84c58a30e6 Cache: fix for sending of empty responses.
Revert wrong fix for empty responses introduced in 0.8.31 and apply new one,
rewritten to match things done by static module as close as possible.
2011-09-27 11:15:35 +00:00
Maxim Dounin
4a3884ae56 Incorrect special case for "return 204" removed.
The special case in question leads to replies without body in
configuration like

    location / { error_page 404 /zero; return 404; }
    location /zero { return 204; }

while replies with empty body are expected per protocol specs.

Correct one will look like

    if (status == NGX_HTTP_NO_CONTENT) {

        rc = ngx_http_send_header(r);

        if (rc == NGX_ERROR || r->header_only) {
            return rc;
        }

        return ngx_http_send_special(r, NGX_HTTP_LAST);
    }

though it looks like it's better to drop this special case at all.
2011-09-27 11:14:02 +00:00
Maxim Dounin
77ca973d11 Fix for "return 202" not discarding body.
Big POST (not fully preread) to a

    location / {
        return 202;
    }

resulted in incorrect behaviour due to "return" code path not calling
ngx_http_discard_request_body().  The same applies to all "return" used
with 2xx/3xx codes except 201 and 204, and to all "return ... text" uses.

Fix is to add ngx_http_discard_request_body() call to ngx_http_send_response()
function where it looks appropriate.  Discard body call from emtpy gif module
removed as it's now redundant.

Reported by Pyry Hakulinen, see
http://mailman.nginx.org/pipermail/nginx/2011-August/028503.html
2011-09-27 11:13:00 +00:00
Maxim Dounin
ad5ef15e08 Fix for double content when return is used in error_page handler.
Test case:

   location / {
       error_page 405 /nope;
       return 405;
   }

   location /nope {
       return 200;
   }

This is expected to return 405 with empty body, but in 0.8.42+ will return
builtin 405 error page as well (though not counted in Content-Length, thus
breaking protocol).

Fix is to use status provided by rewrite script execution in case
it's less than NGX_HTTP_BAD_REQUEST even if r->error_status set.  This
check is in line with one in ngx_http_script_return_code().

Note that this patch also changes behaviour for "return 302 ..." and
"rewrite ... redirect" used as error handler.  E.g.

    location / {
        error_page 405 /redirect;
        return 405;
    }

    location /redirect {
        rewrite ^ http://example.com/;
    }

will actually return redirect to "http://example.com/" instead of builtin 405
error page with meaningless Location header.  This looks like correct change
and it's in line with what happens on e.g. directory redirects in error
handlers.
2011-09-27 11:11:30 +00:00
Maxim Dounin
bfb42ad97c Fix for incorrect 201 replies from dav module.
Replies with 201 code contain body, and we should clearly indicate it's
empty if it's empty.  Before 0.8.32 chunked was explicitly disabled for
201 replies and as a result empty body was indicated by connection close
(not perfect, but worked).  Since 0.8.32 chunked is enabled, and this
causes incorrect responses from dav module when HTTP/1.1 is used: with
"Transfer-Encoding: chunked" but no chunks at all.

Fix is to actually return empty body in special response handler instead
of abusing r->header_only flag.

See here for initial report:
http://mailman.nginx.org/pipermail/nginx-ru/2010-October/037535.html
2011-09-27 11:09:55 +00:00
Maxim Dounin
5c56f20ae7 Proxy: logging levels tuned, double logging fixed. 2011-09-25 20:04:20 +00:00
Maxim Dounin
aa12f3c7f3 Proxy: whitespaces after chunk size allowed.
Whitespaces after chunk size seems to be be allowed by the "implied *LWS" rule
and emitted by some servers.
2011-09-25 20:03:27 +00:00
Maxim Dounin
28b001f897 Upstream: clearing of u->peer.connection on close.
This fixes crashes observed with some 3rd party balancer modules.  Standard
balancer modules (round-robin and ip hash) explicitly set pc->connection
(aka u->peer.connection) to NULL and aren't affected.
2011-09-25 20:00:36 +00:00
Ruslan Ermilov
b1565cf96b Fixed error message. 2011-09-23 20:15:27 +00:00
Ruslan Ermilov
b75e4a6973 Fixed markup. 2011-09-20 13:15:41 +00:00
Igor Sysoev
457f60ae0f release-1.1.4 tag 2011-09-20 11:18:32 +00:00
Igor Sysoev
e70ed7936a nginx-1.1.4-RELEASE 2011-09-20 11:18:24 +00:00
Maxim Dounin
afe2e3d082 Fix of separate pool for upstream connections (r4117).
Pool may not be created if connection was created but rejected in connect()
call.  Make sure to check if it is here before trying to destroy it.
2011-09-20 10:00:46 +00:00
Maxim Dounin
265e51aa0b Fixed loss of chain links in ngx_event_pipe_read_upstream(). 2011-09-20 09:56:35 +00:00
Maxim Dounin
d1b9a0388d Fixed loss of chain links in fastcgi module. 2011-09-20 09:56:05 +00:00
Maxim Dounin
e4dab80e80 Fix of cpu hog in event pipe.
If client closed connection in ngx_event_pipe_write_to_downstream(), buffers
in the "out" chain were lost.  This caused cpu hog if all available buffers
were in the "out" chain.  Fix is to call ngx_chain_update_chains() before
checking return code of output filter to avoid loosing buffers in the "out"
chain.

Note that this situation (all available buffers in the "out" chain) isn't
normal, it should be prevented by busy buffers limit.  Though right now it
may happen with complex protocols like fastcgi.  This should be addressed
separately.
2011-09-20 09:55:27 +00:00
Igor Sysoev
c8df23cb29 The "worker_aio_requests" directive.
The default value is 32 AIO simultaneous requests per worker. Previously
they were hardcoded to 1024, and it was too large, since Linux allocated
them early on io_setup(), but not on request itself. So with default value
of /proc/sys/fs/aio-max-nr equal to 65536 only 64 worker processes could
be run simultaneously. 32 AIO requests are enough for modern disks even if
server runs only 1 worker.
2011-09-20 07:30:09 +00:00
Ruslan Ermilov
a823c550e4 Replaced "can not" with "cannot" and "could not" in a bunch of places.
Fixed nearby grammar errors.
2011-09-19 14:48:29 +00:00
Ruslan Ermilov
3b7ed02583 Cosmetics: replaced NGX_CONF_TAKE1 to NGX_CONF_FLAG for "sendfile"
and "chunked_transfer_encoding" directives, to be in line with all
directives taking a boolean argument.  Both flags will ensure that
a directive takes one argument.
2011-09-19 12:41:13 +00:00
Ruslan Ermilov
5d4aed0175 Overhauled diagnostic messages. 2011-09-19 12:37:22 +00:00
Igor Sysoev
aa997c0289 Fixing Linux AIO initiatialization: AIO operations are disabled if kernel
does not support them.  Previously worker just exited.
2011-09-16 13:41:52 +00:00
Igor Sysoev
4134b48b68 Fixing Linux AIO syscalls return value handling:
syscall(2) uses usual libc convention, it returns -1 on error and
sets errno. Obsolete _syscall(2) returns negative value of error.

Thanks to Hagai Avrahami.
2011-09-16 12:43:16 +00:00
Maxim Dounin
c6353a5654 Trailing space fix. 2011-09-16 12:08:52 +00:00
Maxim Dounin
44002e541f Upstream keepalive module. 2011-09-15 19:28:18 +00:00
Maxim Dounin
2fe0a7a0f6 Proxy: basic HTTP/1.1 support (including keepalive).
By default we still send requests using HTTP/1.0.  This may be changed with
new proxy_http_version directive.
2011-09-15 19:23:20 +00:00
Maxim Dounin
f2fcf11b25 Protocol version parsing in ngx_http_parse_status_line().
Once we know protocol version, set u->headers_in.connection_close to indicate
implicitly assumed connection close with HTTP before 1.1.
2011-09-15 19:22:35 +00:00
Maxim Dounin
f84c69a301 Upstream: Connection header processing. 2011-09-15 19:21:19 +00:00
Maxim Dounin
4686f30a0c Upstream: Transfer-Encoding header processing. 2011-09-15 19:20:08 +00:00
Maxim Dounin
2da36b30c3 Keepalive support in fastcgi.
By default follow the old behaviour, i.e. FASTCGI_KEEP_CONN flag isn't set
in request and application is responsible for closing connection once request
is done.  To keep connections alive fastcgi_keep_conn must be activated.
2011-09-15 19:05:23 +00:00
Maxim Dounin
05552a23a2 Keepalive support in memcached. 2011-09-15 19:04:25 +00:00
Maxim Dounin
2d6be3fe93 Upstream: keepalive flag.
This patch introduces r->upstream->keepalive flag, which is set by protocol
handlers if connection to upstream is in good state and can be kept alive.
2011-09-15 19:03:15 +00:00
Maxim Dounin
a746bab7c1 Upstream: pipe length and input_filter_init in buffered mode.
As long as ngx_event_pipe() has more data read from upstream than specified
in p->length it's passed to input filter even if buffer isn't yet full.  This
allows to process data with known length without relying on connection close
to signal data end.

By default p->length is set to -1 in upstream module, i.e. end of data is
indicated by connection close.  To set it from per-protocol handlers upstream
input_filter_init() now called in buffered mode (as well as in
unbuffered mode).
2011-09-15 19:00:47 +00:00
Maxim Dounin
ffe4f11417 Upstream: r->upstream->length type change to off_t.
Previous use of size_t may cause wierd effects on 32bit platforms with certain
big responses transferred in unbuffered mode.

Nuke "if (size > u->length)" check as it's not usefull anyway (preread
body data isn't subject to this check) and now requires additional check
for u->length being positive.
2011-09-15 18:43:19 +00:00
Maxim Dounin
e19f005daf Upstream: content_length_n API change.
We no longer use r->headers_out.content_length_n as a primary source of
backend's response length.  Instead we parse response length to
u->headers_in.content_length_n and copy to r->headers_out.content_length_n
when needed.
2011-09-15 18:33:43 +00:00
Maxim Dounin
360ed25d65 Upstream: separate pool for peer connections.
This is required to support persistent https connections as various ssl
structures are allocated from connection's pool.
2011-09-15 18:21:24 +00:00
Maxim Dounin
c42c70f478 Workaround for cpu hog on errors with cached connections.
Just doing another connect isn't safe as peer.get() may expect peer.tries
to be strictly positive (this is the case e.g. with round robin with multiple
upstream servers).  Increment peer.tries to at least avoid cpu hog in
round robin balancer (with the patch alert will be seen instead).

This is not enough to fully address the problem though, hence TODO.  We
should be able to inform balancer that the error wasn't considered fatal
and it may make sense to retry the same peer.
2011-09-15 18:12:58 +00:00
Maxim Dounin
d7c2673d3f API change: ngx_chain_update_chains() now requires pool.
The ngx_chain_update_chains() needs pool to free chain links used for buffers
with non-matching tags.  Providing one helps to reduce memory consumption
for long-lived requests.
2011-09-15 16:03:17 +00:00
Maxim Dounin
a890b313f3 Buffers reuse in chunked filter.
There were 2 buffers allocated on each buffer chain sent through chunked
filter (one buffer for chunk size, another one for trailing CRLF, about
120 bytes in total on 32-bit platforms).  This resulted in large memory
consumption with long-lived requests sending many buffer chains.  Usual
example of problematic scenario is streaming though proxy with
proxy_buffering set to off.

Introduced buffers reuse reduces memory consumption in the above problematic
scenario.

See here for initial report:
http://mailman.nginx.org/pipermail/nginx/2010-April/019814.html
2011-09-15 15:25:42 +00:00