On Tue, Feb 11, 2014 at 09:57:55PM -0800, Piotr Sikora wrote:
> # HG changeset patch
> # User Piotr Sikora
> # Date 1392184482 28800
> # Tue Feb 11 21:54:42 2014 -0800
> # Node ID 3f21f5b57864af853ad097a113e419086a603aae
> # Parent cff36d2d7fe6db1baa9d44ed30ebd26b20c05d06
> Upstream: fix $up
Hello!
On Wed, Feb 12, 2014 at 12:18:55PM +0400, Ruslan Ermilov wrote:
> On Tue, Feb 11, 2014 at 09:57:55PM -0800, Piotr Sikora wrote:
> > # HG changeset patch
> > # User Piotr Sikora
> > # Date 1392184482 28800
> > # Tue Feb 11 21:54:42 2014 -0800
> > # Node ID 3f21f5b57864af853ad097a113e4
details: http://hg.nginx.org/nginx/rev/d15822784cf9
branches:
changeset: 5575:d15822784cf9
user: Piotr Sikora
date: Tue Feb 11 21:54:42 2014 -0800
description:
Upstream: fix $upstream_status variable.
Previously, upstream's status code was overwritten with
cached response's status c
Hello!
On Tue, Feb 11, 2014 at 01:16:41PM -0800, Piotr Sikora wrote:
> Hey Maxim,
>
> > Well, there is no real difference, but I think that it would be
> > easier to use distinct flags instead. Note that it also matches
> > what Apache has:
> >
> > http://httpd.apache.org/docs/2.4/mod/mod_ssl.h
details: http://hg.nginx.org/nginx/rev/2bc609a4b516
branches:
changeset: 5576:2bc609a4b516
user: Valentin Bartenev
date: Wed Feb 12 21:02:29 2014 +0400
description:
SPDY: fixed reversed priority order in window waiting queue.
diffstat:
src/http/ngx_http_spdy.h | 3 +
On Wednesday 12 February 2014 20:29:08 Maxim Dounin wrote:
[..]
> > > proxy_ssl_sni on|off
> > >
> > > default: off (?)
> > > flag, controls if SNI (Server Name Indication) will be used
> > > while connecting to backends;
> > >
> > > (I tend to think that "proxy_ssl_sni" is a better nam
Hi,
I am maintaining a custom module for nginx. We currently rotate our logs
and load dynamic data every 5 minutes by using SIGHUP. Unfortunately this
is also tears down our keepalive connections which harms our performance
(we are handling about 1500 requests per second per server). We have a
num
