Hey Jürgen,

can you provide us with some more information on your instance. Especially, is there any log output in /var/log/openproject/production.log that would be of any use?

I couldn’t reproduce the error on an updated instance on 14.04LTS. Since you wrote you’re using nginx, can you provide us with the configuration changes since the package doesn’t natively support it?

Best,
Oliver

Hello Oliver,

Thanks for your help.

The log files you can find at: https://dl.dropboxusercontent.com/u/55490396/var-log-openproject.zip

The OpenProject server is behind a reverse proxy (Ubuntu 14.04LTS, NGINX). Config on that server:
————————————————————————————————————————————————————————————
server {
listen 80;
server_name project.jrweiss.com;
rewrite ^/(.*) https://$server_name/$1 permanent; # force redirect http to https
# return 301 https://$server_name/$1; # force redirect http to https
}

server {
listen 443;
server_name project.jrweiss.com; # OpenProject on weiss-lms-109

ssl on;
ssl_certificate /etc/ssl/nginx/wildcard.jrweiss.com.crt; # path to certificate
ssl_certificate_key /etc/ssl/nginx/wildcard.jrweiss.com.key; # path to key

root /var/www/html;
index index.html index.htm index.php index.nginx-debian.html;

access_log /var/log/nginx/project.access.log;
error_log /var/log/nginx/project.error.log;

location / {
proxy_pass http://10.0.23.109:88; # OpenProject

# force timeouts if one of backend died ##
proxy_next_upstream error timeout invalid_header http_500 http_502 http_503 http_504;

proxy_set_header Accept-Encoding “”;
proxy_set_header Host $host;
proxy_set_header X-Real-IP $remote_addr;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_set_header X-Forwarded-Proto $scheme;
# add_header Front-End-Https on;
}
}
————————————————————————————————————————————————————

The NGINX confile file on the actual OpenProject server:
————————————————————————————————————————————————————
passenger_root /usr/lib/ruby/vendor_ruby/phusion_passenger/locations.ini;
passenger_ruby /usr/bin/passenger_free_ruby;
passenger_max_pool_size 10;

server {
listen 88;
server_name project.jrweiss.com 10.0.23.109;
passenger_enabled on;
passenger_friendly_error_pages on;

#Path to rails app’s public directory
root /var/opt/openproject/public;

access_log /var/log/nginx/project.jrweiss.com-access.log;
error_log /var/log/nginx/project.jrweiss.com-error.log;

include global/restrictions.conf;

location ~ ^/(.*|$) {
proxy_pass_request_headers on;
proxy_set_header X-Forwarded-Host $host;
proxy_set_header X-Forwarded-Server $host;
proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
proxy_pass http://127.0.0.1:6000;
}
}
—————————————————————————————————————————————-

Content of restrictions.conf:
——————————————————————————————————————————————

# Global restrictions configuration file.

# Designed to be included in any server {} block.

location = /favicon.ico {
log_not_found off;
access_log off;
}

location = /robots.txt {
allow all;
log_not_found off;
access_log off;
}

1. Deny all attempts to access hidden files such as .htaccess, .htpasswd, .DS_Store (Mac).
2. Keep logging the requests to parse later (or to pass to firewall utilities such as fail2ban)
   location ~ /\. {
   deny all;
   }

1. Deny access to any files with a .php extension in the uploads directory
2. Works in sub-directory installs and also in multisite network
3. Keep logging the requests to parse later (or to pass to firewall utilities such as fail2ban)
   location ~* /(?:uploads|files)/.*\.php$ {
   deny all;
   }
   ———————————————————————————————————————————————————————

Hi

no one that can help? :-(

Thanks.

Hello,

Looks like it it is a OpenProject bug.

1. I created a new server with a fresh OpenProject installation using Apache,
2. Copied the content of the files and config directories from the existing to the new server and imported the database. - Same behaviour on new server like on old server.
3. Then did above with the config from the new server, only the database import but did not copy config from old server over. - Same behaviour.

Seems whatever is in the database causes the problem

Now I am really stuck, don’t want to re-input the data which is on the old server.

Any ideas?

Thanks.

Hey Jürgen,

I can’t really provide any help to the nginx configuration. It looks good to me, but we don’t officially support that method.

Judging from the production log, there appears to be a missing setting due to the custom installation step (that setting should have been set while configuring the apache addon of the installer).

I see the following errors:

E, [2016-05-30T19:21:42.126420 #4385] ERROR -- : 2016-05-30T19:21:42+0100: [Worker(host:weiss-lms-109 pid:4385)] Job DeliverWorkPackageNotificationJob (id=21) FAILED (9 prior attempts) with ActionView::Template::Error: Invalid :protocol option: ""

-> Please set the ‘Protocol’ setting in the global settings (Administration > System Settings). This should usually be taken care of by the installer, but probably originates from the Apache setup.

The second error stems from the file /opt/openproject/config/configuration.yml, which contains some ERB interpolated values to configure the application from ENV variables. (That’s what happens internally when you configure the application with openproject config:set).

I, [2016-04-05T10:44:25.409975 #7107]  INFO -- : Refreshing Gem list
/opt/openproject/vendor/ruby-2.1.6/lib/ruby/2.1.0/erb.rb:850:in `eval': (erb):40: syntax error, unexpected ':', expecting ')' (SyntaxError)
  attachments_storage_path: <%= ENV.fetch('ATTACHMENTS_S...
                           ^
(erb):41: syntax error, unexpected tCONSTANT, expecting ')'
;  git_configured = ENV['GIT_REPOSITORIES'].present?...
                       ^
(erb):62: syntax error, unexpected end-of-input, expecting ')'
; _erbout.force_encoding(__ENCODING__)
                                      ^
    from /opt/openproject/vendor/ruby-2.1.6/lib/ruby/2.1.0/erb.rb:850:in `result'
    from /opt/openproject/lib/open_project/configuration.rb:227:in `load_config_from_file'

Did you change something in that file? If so, can you please post that file?

Best,
Oliver

Hello Oliver,

thank you for coming back, you helped me with above to fix it.

It was not a protocol issue as you suggested above, that one was set correctly to HTTP (but it helped me to find the problem :-)).

To fix the problem (unable to edit work packages, no split screen, etc.) one needs to reset the “Host name” value in “Administration - System settings”. Although it was still set correctly after the upgrade. I simply input the same value again, e.g. “project.domain.com” (plus saved it) and voila all working again.

As per my previous reply I did a clean OpenProject/Apache install and after importing the database had the same problem. So, it is not a NGINX issue.

Re. the second error, I did not change the /opt/openproject/config/configuration.yml file, but will have a look.

Many, many thanks. You saved me a lot of work, was already prepared to redo the install and retype all data.

Have a great day. regards, Jürgen

Hey Jürgen,

great to hear! I’ll see that we extend the installation guide with a notice that some values need to be set manually when diverting from the default installation path (Hostname should also be set from within Apache),
so that the necessary manual steps are easier to understand.

Let me know if the issue with the configuration.yml persists.

Best,
Oliver