Hi,
I have successfully reverse proxied some 180 applications. Due to CGI constraints, I am listening on only one port, and have created separate location blocks for each application, with sub filters modifying the urls with redirects and sub_filters.
While performance is no issue, on restarting nginx, The operation frequently times out, as so many applications are looked up. On production servers, it causes a big issue. As far as possible, I reload instead of restart and test the configuration with the -c and -t options, The restart still times out sometimes.
I have to restart because sometimes, the proxy_pass hosts' DNS entries change, and I haven't added a variable to proxy_pass with an upstream {} DNS.
The question is this: Can I restart nginx without it having to look up every upstream app server when restarting? How can I prevent pre configuration operations from timing out or sending fatal signals to the restart process?
I have successfully reverse proxied some 180 applications. Due to CGI constraints, I am listening on only one port, and have created separate location blocks for each application, with sub filters modifying the urls with redirects and sub_filters.
While performance is no issue, on restarting nginx, The operation frequently times out, as so many applications are looked up. On production servers, it causes a big issue. As far as possible, I reload instead of restart and test the configuration with the -c and -t options, The restart still times out sometimes.
I have to restart because sometimes, the proxy_pass hosts' DNS entries change, and I haven't added a variable to proxy_pass with an upstream {} DNS.
The question is this: Can I restart nginx without it having to look up every upstream app server when restarting? How can I prevent pre configuration operations from timing out or sending fatal signals to the restart process?