On Fri, Jan 23, 2009 at 02:55:03PM -0500, John Kemp wrote:
On Jan 23, 2009, at 2:09 PM, Igor Sysoev wrote:
On Fri, Jan 23, 2009 at 12:56:14PM -0500, John Kemp wrote:
I am using nginx in a reverse-proxy situation, with several upstream
servers on different networks.
I am often testing parts of my server without access to some of the
upstream servers. When I try to start nginx and even a single
upstream
server is unavailable, I get the message:
comp:blah johnk$ sudo /usr/local/sbin/nginx
2009/01/23 10:24:22 [emerg] 4894#0: host not found in upstream
"xxx.yyy.zzz" in /usr/local/nginx//conf/nginx.conf:18
And nginx is not started.
Although I like configuration checking, I'd also like to be able to
start nginx even when one or more upstream servers are down. Is there
a way to do that? I couldn't see a command-line option to do so, or
anything that looked relevant in the config file examples on the
wiki.
I am using nginx 0.7.30, compiled from source on OS X.
One of the way - to use IP addresses.
Is there a way to disable on-startup host lookup altogether (other
than making a code change)?
There is complex way - using variables in proxy_pass:
proxy_pass http://host$request_uri;
but you can not disable lookups for servers inside upstream blocks.
4 messages in ru.sysoev.nginxRe: Starting nginx without checking f...
