I am trying to configure Nginx to proxy pass to Gunicorn.
My django project can be found at /home/justin/project/jobzumo
Start by creating and opening a new server block in Nginx’s sites-available directory:
sudo nano /etc/nginx/sites-available/jobzumo
Within this file I've entered the following:
server{
listen 80;
server_name 142.93.184.125;
location = /favicon.ico { access_log off; log_not_found off; }
location /static/ {
root /home/justin/project;
}
location / {
include proxy_params;
proxy_pass http://unix:/run/gunicorn.sock;
}
}
When I go to 142.93.184.125 I see the default django rocket ship, so I think that means everything is working. However, when I go to 'jobzumo.com' (associated domain), I see the default 'Welcome to nginx!' page.
I know I have both the IP and domain name in my ALLOWED_HOSTS settings and have pointed the domain nameservers at my server. So, do I need to add this domain to this file? The tutorial I was following said either or should do the job. If adding the domain to this file is not what I have to do, can you mention that, so at least I know I'll have to start looking elsewhere. Thanks for any help.
You probably still have the default site in available sites in nginx which is causing the issue. I just had the same problem and the following two commands solved the issue:
sudo unlink /etc/nginx/sites-enabled/default
sudo service nginx restart
if you stopped your gunicorn daemon you might need to restart it and then run the second command above it should do the trick.
Related
I'm using Docker to serve my simple WordPress website. A nginx container and a wordpress container. Simple setup:
upstream wordpress_english {
server wordpress_en:80;
}
server {
listen 80;
server_name my_domain.com www.my_domain.com;
location / {
proxy_pass http://wordpress_english;
}
}
Problem: Static files (css, js and images) are not loaded.
The output from the browser console shows a 404:
http://wordpress_english/wp-content/themes/twentyfifteen/genericons/genericons.css?ver=3.2
Its easy to spot the problem: The browser looks for the static files at wordpress_english (the nginx upstream name), instead of my_domain.com
How can I fix this problem?
This is not a nginx problem, but a WordPress problem.
Solution:
In wp-config.php, add the following two lines:
define('WP_HOME','http://my_domain.com');
define('WP_SITEURL','http://my_domain.com');
During the WordPress installation, WordPress automatically sets WP_HOME to nginx upstream name. The above solution overwrites the default setting.
Seems to be an issue in your nginx config file.
When declaring your server my_domain you provide location / with proxy_pass wordpress_english. I don't know a lot on nginx, but I don't see any declaration of path in your server my_domain and is root is linked to wordpress_english. Seems normal that he is looking for files in wordpress_english and not in you server. (In fact, I guess he is looking in your server but your server tells to look in wordpress).
Not sure about it cause I don't know well nginx and proxy_pass functions.
With the Synology DSM6 update, we have now to use Nginx instead of Apache. By default Nginx configuration don't allow wordpress permalinks (generate 404).
I read the idea was to transform the /uri in /?p=$uri and put this configuration in the "location" section of the server nginx config.
Where to put this configuration in DSM6 exactly ?
Have you tried the user config? Just copy your working:
/etc/nginx/app.d/server.webstation-vhost.conf
to:
/usr/local/etc/nginx/sites-enabled/httpd-vhost.conf-user
and rename the server.webstation-vhost.conf to server.webstation-vhost.conf.old or something and restart nginx (nginx -s reload)
Or better yet, remove your virtual host(s) from webstation. Only thing is you need to manually update your SSL certs when they expire instead of using the web interface.
Actually, you can add custom directives easily, without modifying the DSM behavior.
Take a look at the content of /usr/local/etc/nginx/sites-enabled/httpd-vhost.conf-user, to see where the custom configuration has to be stored:
server {
[...]
server_name NAME
[...]
include /usr/local/etc/nginx/conf.d/778943ad-0dc4-40ae-bb7f-7b2285e3203b/user.conf*;
}
Then, you just have to create the file /usr/local/etc/nginx/conf.d/778943ad-0dc4-40ae-bb7f-7b2285e3203b/user.conf.wordpress-permalink with the following content:
location /{
try_files $uri $uri/ /index.php?$args;
}
and restart nginx:
synoservicecfg --restart nginx
It will not break the future DSM update (since it is a supported customization)
I have a Play application listening on a local port :9000. There are other applications running.
I would like to server this application at a path like:
http://myhost/this-play-app -> localhost:9000
So that other apps could be nested at other paths.
I've tried the basic proxy_pass but it doesn't seem to work.
server {
listen 80;
server_name myhost;
# MMC Tool
# ----------------------------------------------------
location /this-play-app {
proxy_pass http://localhost:9000;
}
}
The play app seems to forward to the root. Is there a way to trick the play app to work within the /this-play-app path ?
Like /this-play-app/some-controller instead of /some-controller ?
Thanks
Using apps in folders isn't comfortable idea - you would need to at least prepare some dedicated config and change it each time when changing the location.
Instead as other suggested you should use subdomains, in this case each app behaves exactly the same as in the root domain and even if you will need/want to change that domain all you'll need will be change in the nginx's config.
Typical nginx's config looks like
upstream your_app {
server 127.0.0.1:9000;
}
server {
listen 80;
server_name your-app.domain.com;
location / {
proxy_pass http://your_app;
}
}
Most probably on some VPS or shared hosts you'll need to add the subdomain by some kind of admin's panel - on localhost just need add the subdomain to the hosts file.
Edit if using subdomain is not possible anyway (pity) you can workaround it anyway by config, in nginx use (as you did in question:
...
location /this-play-app {
proxy_pass http://your_app;
}
...
and then add this line into your application.conf (Play 2.1+)
application.context = "/this-play-app"
Or this in case of Play 2.4+ (info)
play.http.context = "/this-play-app"
I want to test nginx subdomains before uploading config to the server. Can i test it on localhost? I try
server {
listen 80;
server_name localhost;
location / {
proxy_pass http://localhost:8080;
}
}
server {
listen 80;
server_name sub.localhost;
location / {
proxy_pass http://localhost:8080/sub;
}
}
And it does not work. Shoulld i change my hosts file in order to make it work? Also, after uploading site to the server should i change DNS records and add sub.mydomain.com?
Yes, add '127.0.0.1 sub.localhost' to your hosts file. That sub has to be resolved somehow. That should work.
Then once you're ready to go to the net, yes, add an a or cname record for the subdomain sub.
When I use proxy_pass I also include the proxy.conf from nginx.
http://wiki.nginx.org/HttpProxyModule
In Linux based OS just to edit as sudo /etc/hosts file and change 127.0.0.1 localhost to 127.0.0.1 *.localhost.
So at /etc/nginx/sites-enabled/<environment>/<your_project_name> edit server_name key as <subdomain>.localhost.
Reload nginx and networking service.
$ sudo service nginx reload
$ sudo service networking reload
And then try http://<subdomain>.localhost at url bar.
It works for me.
UPDATE
In my opinion, a better solution is creating a virtual server that only responds if subdomain doesn’t exist, at /etc/nginx/sites-enabled/development/default, as default server (remember that you can define only one server as default).
server {
listen 80 default_server;
root /var/www/html/errors/404;
server_name *.localhost *.<host-name>;
location / {
index subdomain.html;
}
}
Make sure that in nginx.conf (generally at /etc/nginx/nginx.conf) contain include /etc/nginx/sites-enabled/**/*; to this virtual server work. If not, put it and then run $ sudo service nginx reload.
In this case isn't necessary put *.localhost in /etc/hosts, but only localhost.
For your public webserver with its own domain name, you just need to add a Canonical name using a CNAME record in your DNS configuration:
CNAME * example.com.
Once this is done, set your nginx setting
server_name *.example.com example.com;
In your local setup you can keep the same configuration for nginx but unless you have a local DNS setup, you will have to edit your /etc/hosts file and add each subdomain manually. wildcards don't work in the /etc/hosts file.
127.0.0.1 abc.example.com def.example.com ghi.example.com
It is generally recommended to use .local as the namespace for your local domains.
With an Nginx configuration like shown by the OP, all that is needed is to configure the local DNS resolution. I run Linux containers on a VM with a local DHCP IP but test them on Windows 10 browsers.
The DNS configuration can be done by editing "C:\Windows\System32\drivers\etc\hosts" as Administrator.
192.168.100.50 sub.example.local
192.168.100.50 example.local
Of course, use 127.0.0.1 or other appropriate IP as needed.
I got a new slice off slicehost, for the purposes of playing around and learning nginx and more about deployment generally. I installed a ruby app on there (which i'll call app1) which uses passenger. I made it the default app to use for that server with the following server block in my nginx config:
server {
listen 80;
server_name <my server ip>;
root <path to app1 public folder>;
passenger_enabled on;
}
This works fine. However, i want to try a few different apps out on this slice, and so thought i would set it up like so:
http:///app1
http:///app2
etc. I thought i would be able to do that by adding a location block, and moving the app1 specific stuff into it like so:
server {
listen 80;
server_name <my server ip>;
location ^~ /app1 {
root <path to app1 public folder>;
passenger_enabled on;
}
}
However, on doing this (and restarting nginx of course), going to the plain ip address gives the 'welcome to nginx' message (which i'd expect). But, going to /app1 gives an error message:
404 Not Found
The requested URL /app1 was not found on this server.
This is distinct from the error message i get when i go to another path on that ip, eg /foo:
404 Not Found
nginx/0.8.53
So, it's like nginx knows about that location but i've not set it up properly. Can anyone set me straight? Should i set up different server blocks instead of using locations? I'm sure this is simple but can't work it out.
Cheers, max
What you're after is name virtual hosting. The idea is that each domain is hosted on the same IP, and nginx chooses the virtualhost to serve based on the Host: header in the HTTP request, which is sent by the browser.
To use name virtual hosting, use the domain you want to serve instead of your server's IP for the server_name directive.
server {
listen 80;
server_name app1.com;
location / {
root /srv/http/app1/public;
passenger_enabled on;
}
}
Then, to host more apps on the same box, just declare a separate server { } block for each one.
server {
listen 80;
server_name app2.com;
location / {
root /srv/http/app2/public;
passenger_enabled on;
}
}
I'm using unicorn instead of passenger, but the vhost part of the structure is the same for any backend.
The global nginx config (which on its own hosts nothing): https://github.com/benhoskings/babushka-deps/blob/master/nginx/nginx.conf.erb
The template wrapper for each virtualhost: https://github.com/benhoskings/babushka-deps/blob/master/nginx/vhost.conf.erb
The details of the unicorn virtualhost: https://github.com/benhoskings/babushka-deps/blob/master/nginx/unicorn_vhost.common.erb
I fail to see the real problem here tho,
in order for you to figure that out
you need to view the nginx log files on most systems at:
/var/log/nginx/
and open the relevant access file here(might be error.log)
in there you can see what url nginx exactly tried to access and why did it fail.
What I really think is happening, that you got the root path wrong,
maybe it should be alias instead because
if you are proxifying the connection to another app, it might get the
"app1" word in the url instead of a direct one.
so please try:
server {
listen 80;
server_name <my server ip>;
location /app1 {
alias <path to app1 public folder>;
passenger_enabled on;
}
}
and see weather it works and also try to view the logs first to really determine whats the problem.
I think its just a slight syntax problem:
location ~ ^/app1 { ...
should work, or a little more efficient:
location = /app1 { ...
One problem is that your Rails app probably wasn't designed to run from a subdirectory. Passenger has a directive that will fix this:
passenger_base_uri /app1;
However, running Rails apps in subdirectories is somewhat non-standard. If you can, a better option may be to set up subdomains using nginx's virtual hosts.
It seems that you want to host more apps on the same server with base uri. Try this:
root /srv/http/;
passenger_base_uri /app_1;
passenger_base_uri /app_2
Also under /srv/http, create 2 symlinks:
ln -s /srv/http/app_1 /srv/http/app1/public
ln -s /srv/http/app_2 /srv/http/app2/public
The app1 can be accessed under: http://domain.com/app_1.
Here is more for reading: http://www.modrails.com/documentation/Users%20guide%20Nginx.html#deploying_rack_to_sub_uri