I am trying to get passenger to work but i keep seeing the apache default page instead. I am using centos 6.3 on an inmotion VPS. Here is what I've done so far:
gem install passenger.
rvmsudo passenger-install-apache2-module
pasted the following code into /usr/local/apache/conf/includes/post_virtualhost_global.conf (this gets loaded into httpd.conf)
LoadModule passenger_module /home/username/.rvm/gems/ruby-1.9.3-p194/gems/passenger-3.0.17/ext/apache2/mod_passenger.so
PassengerRoot /home/username/.rvm/gems/ruby-1.9.3-p194/gems/passenger-3.0.17
PassengerRuby /home/username/.rvm/wrappers/ruby-1.9.3-p194/ruby
<VirtualHost 123.123.123.123:80>
ServerName http://XXXXXX.inmotionhosting.com
# !!! Be sure to point DocumentRoot to 'public'!
DocumentRoot /home/username/rails/current/public
<Directory /home/username/rails/current/public>
# This relaxes Apache security settings.
AllowOverride all
# MultiViews must be turned off.
Options -MultiViews
</Directory>
</VirtualHost>
edit: not sure why this isnt formatting properly. its lined and tabbed fine in the text editor.
in etc/hosts there is the line:
123.123.123.123 XXXXXX.inmotionhosting.com XXXXXX
i restarted apache via sudo service httpd restart in /etc/init.d
still shows apache default page
Any ideas? thanks
I stumbled upon this 1 year old post for the similar problem but none of the comment worked for me. So I'm just adding my 2 cents for those who are looking for some help.
If you are seeing Apache default page, it means that phusion passenger is probably not running. You might have seen a warning message when you started/stopped/restarted the apache server that says the phusion passenger module is already loaded, skipping.
So Apache has skipped the most recent LoadModule line you added in the conf file.
Then check the error.log which is in the directory set for error log in apache config file. (httpd.conf for centos)
In my case, it was complaining that it couldn't find the watchdog in the directory set in PassengerRuby. This confirms that the passenger module loaded is NOT what I just added. This might have happened because I am using RVM now but also installed an old version of Passenger a while ago as a root. So I looked at the config.d directory under Apache root, there was a file "passenger.conf" that was loading the old version of Passenger.
I deleted the file, kept the new LoadModule as is - copied from the instruction Phusion Passenger gave me while installing apache2-module, restarted the apache server, no warning message, and passenger worked.
Related
I am trying to load a NodeJS App from Apache+Passenger.
OS: Centos 7 / CloudLinux release 7.9
Apache: 2.4.6 ( httpd-2.4.6-97.el7_9.cloudlinux.x86_64 )
Passenger: Phusion Passenger 6.0.7 ( passenger-6.0.7-1.el7.x86_64, mod_passenger-6.0.7-1.el7.x86_64 ) from Passenger's Yum repo
The virtual host points to /home/vhost1/public_html and the NodeJS App points to /home/vhost1/nodeapps/np1-pass/np1-pass.js
When I run this config via Passenger entries in Apache's config, the application works and is accessible from http://virtual-host/np1-pass/ .
The entry in Apache's config is as under :
<VirtualHost *:80>
ServerName virtual-host
DocumentRoot /home/vhost1/public_html
<Directory /home/vhost1>
# Relax Apache security settings
AllowOverride all
Require all granted
# MultiViews must be turned off
Options -MultiViews
</Directory>
<Directory /home/vhost1/public_html>
# Relax Apache security settings
AllowOverride all
Require all granted
# MultiViews must be turned off
Options -MultiViews
</Directory>
Alias /np1-pass /home/vhost1/nodeapps/np1-pass/public
<Location /np1-pass>
PassengerAppEnv development
Passengerapproot /home/vhost1/nodeapps/np1-pass
PassengerBaseURI "/np1-pass"
PassengerNodejs "/home/vhost1/bin/node"
PassengerAppType node
PassengerStartupFile np1-pass.js
</Location>
</VirtualHost>
What now I have been trying was to move over Passenger directives to the .htaccess file under /home/vhost1/public_html/np1-pass (of-course after removing those directives and the alias and location entry from the Apache config) so that a virtual host owner is able to fire the application without modifying Apache's config, then I get the error
PassengerAppRoot not allowed here, referer: http://virtual-host/np1-pass/
A similar .htaccess works fine for CloudLinux 7 on a cPanel server with Passenger ea-apache24-mod-alt-passenger-5.3.7-9.el7.cloudlinux.x86_64 package installed
When I was looking into various solutions, I came across two documents
https://www.phusionpassenger.com/docs/references/config_reference/apache/#passengerapproot
https://www.phusionpassenger.com/library/config/apache/reference/#passengerapproot
W.r.t. Passenger directives in the 1st link document, under the directive's context .htaccess is missing . But same is present in the 2nd link document
Now I have 2 questions
Are both documents correct or am I missing something. ?
And why the same setup is working fine on CloudLinux 7 on a cPanel server with Passenger ea-apache24-mod-alt-passenger-5.3.7-9.el7.cloudlinux.x86_64 package installed ?
Thanks
Kirti Singh
W.r.t. the question
Are both documents correct or am I missing something. ?
I can't say about that, but I found this useful link https://blog.phusion.nl/2018/01/29/passenger-5-2-0/ which states that many Passenger .htaccess directives have been disabled since Passenger 5.2.0 . But still as both links mentioned in the original question do exist, those links should have had a version prefixed or there should have been a clarification in their content.
Now w.r.t. the 2nd question
And why the same setup is working fine on CloudLinux 7 on a cPanel server with Passenger ea-apache24-mod-alt-passenger-5.3.7-9.el7.cloudlinux.x86_64 package installed ?
I can't say how CloudLinux is doing it, but I was able to run passenger directives via .htaccess by uninstalling the passenger 6.0.7 and the corresponding mod_passenger , and then by installing the old passenger version 5.1 and its corresponding mod_passenger
`yum remove passenger
yum install passenger-5.1.12-1.el7
yum install mod_passenger-5.1.12-1.el7`
This solved my problem and now I am able to load NodeJS apps via .htaccess
I hope my answer helps others who struggle with this issue.
i searched sth about this a lot but solutions doesnt help so much.I tried to upgrade redmine to 2.6.5 on my FreeBSD but i have 403 error.
apache error log:
[autoindex:error] AH01276: Cannot serve directory
/usr/local/www/redmine/public/: No matching DirectoryIndex (none)
found, and server-generated directory index forbidden by Options
directive
my httpd conf:
<VirtualHost example.com:80>
DocumentRoot "/usr/local/www/redmine/public"
ServerName example.com
FastCgiServer /usr/local/www/redmine/public/dispatch.fcgi -idle-timeout 120 -initial-env RAILS_ENV=production -initial-env PATH=/usr/local/bin -processes 2
<Directory "/usr/local/www/redmine/public">
AddHandler fastcgi-script fcgi
Order allow,deny
Allow from all
AllowOverride all
Options +FollowSymLinks +ExecCGI
RewriteEngine On
RewriteCond %{REQUEST_FILENAME} !-f
RewriteRule ^(.*)$ dispatch.fcgi
</Directory>
ErrorLog /logs/error.log
</Virtualhost>
i have to say: if i add +Indexes in Option i see files in my browser, so i guess premissions are good.
Anyone can give me any hint?
thanks in advance 4 your help
In my situation the problem was caused by a bug in the apache module ModAutoIndex. Disabling the module did the trick.
See https://serverfault.com/a/731859
Disabling the module auto index (which is the cause of the wrong behaviour, will prevent the error.
#LoadModule autoindex_module modules/mod_autoindex.so
Phusion will address the issue in the realase of Passenger 5.0.22 before Apache 2.5.0 will be released.
I suffered the same issue on my ArchLinux installation with Apache 2.4 and Redmine 2.6.5. Instead of fcgi I use the Phusion Passenger web application server but I always ended up on a 403 Forbidden page like you when accessing the server.
With the +Indexes option I also got the contents of the Redmine public directory on the browser.
When using webrick or passenger directly to host Redmine everything was fine. That's how you can verify your Redmine is not corrupted somehow. From the Redmine root directory run:
bundle exec ruby scripts/rails server -e production
Since I'm using RVM to manage ruby versions and gemsets on the system I can also tell you that the behavior is not related to ruby ( I tried every version from 1.8.x to 2.2.x without any change ).
In the end I substituted Apache with nginx 1.8.0 (stable release) and got Redmine back to work. So with passenger it's quite easy to get rolling. Just run
gem install passenger
so you get the passenger package. And then compile nginx with the passenger module using
passenger-install-nginx-module
You'll get an automated dialog that downloads nginx and compiles it with the appropriate config. By default it will be installed to /opt/nginx
On ArchLinux you would rather use the ABS to get the PKGBUILD and add the following to the configure part
--add-module=$(passenger-config --nginx-addon-dir) \
That way you also get the systemd unit-file to start and stop nginx.service
Last but not least here's the nginx config I use to run Redmine:
server {
listen 80;
server_name redmine.example;
root /usr/share/webapps/redmine-2.6.5/public;
passenger_base_uri /;
passenger_app_root /usr/share/webapps/redmine-2.6.5;
passenger_document_root /usr/share/webapps/redmine-2.6.5/public;
passenger_enabled on;
passenger_ruby /usr/local/rvm/gems/ruby-1.9.3-p551#redmine/wrappers/ruby;
}
Using another webserver may be a daunting step but it took me less than 2 hours to get Redmine up and running with nginx compared to nearly 2 days of wasted time to figure out why the heck Apache didn't serve the webapp any more.
I have upgraded to Apache 2.2.22 on a Debian 7 VM. I have just installed it using apt-get and not made any config changes except to add my VirtualHost
<VirtualHost *:8001 *:80>
ServerName devintegration.ie
ServerAlias devintegration.ie
DocumentRoot /sites/integration/development/docs
ScriptAlias /console/ /sites/integration/development/console/
ErrorLog /sites/logs/devintegration.ie-error_log
CustomLog /sites/logs/devintegration.ie-access_log common
</VirtualHost>
The behavior I was used to was that any Perl errors or warns appeared in the error log specified in the virtual host. The behavior I'm now getting is that the Perl errors go to the common Apache log in /var/log/apache2 directory and the virtual host log just contains "Premature end of script headers".
I've done some googling on this but most of the results are about mod_perl. I did find this page and the last comment suggests that there was a change to how this works but no information was given as to whether it could be changed. Apache Log Forum "Any information written to stderr by a CGI script will be copied directly to the error log."
Is there a way to configure apache to direct Perl errors and warns to the error log specified in virtual host?
The previous Apache version I was using was 2.2.15.
This is (was?) a longstanding bug in mod_cgid. Its status still is "RESOLVED LATER" which (apparently) means: "stop complaining - we'll do it in a later version, not now". Maybe it has been solved by now (your version, 2.2.22, is 3 years old), but I couldn't find any mention of it.
mod_cgid is the default .cgi handler in the threaded version of Apache. If you switch to mod_cgi, e.g. by using a non-threaded apache (like I do) that should solve your problem.
I am trying to use Passenger to deploy a rails app on Ubuntu with Apache. I've followed pretty much every guide I could find on configuring Passenger with Apache/Ubuntu. However, when I go to the site, I just see the actual file directories and not the running version of the application. If i switch to the /public directory I just see the public directory when I go to the site. I can start and stop the rails server by calling rails s without getting any errors, but still am not able to see the running version of the application. Does anyone ran into similar problems? Is there a good starting point anyone could suggest I look at? I'm not sure if the problem is with apache or passenger.
Apache config
LoadModule passenger_module /usr/lib/ruby/gems/1.9.1/gems/passenger-3.0.9/ext/apache2/mod_passenger.so
PassengerRuby /usr/bin/ruby
PassengerRoot /usr/lib/ruby/gems/1.9.1/gems/passenger-3.0.9
PassengerMaxPoolSize 10
<VirtualHost *:80>
ServerName http://www.myapp.com/
DocumentRoot /var/www/myapp/current/
<Directory /var/www/myapp/current/public>
Allow from all
</Directory>
When I restart apache, I get a message stating [Sun Nov 06 14:43:48 2011] [warn] module passenger_module is already loaded, skipping I'm not sure if this is an error or is normal.
Update
When I run find / -name 'mod_passenger.so' I get two locations:
/usr/lib/ruby/gems/1.9.1/gems/passenger-3.0.9/ext/apache2/mod_passenger.so
/usr/lib/apache2/modules/mod_passenger.so
Update 2
I have several passenger.config files
/etc/apache2/mods-enabled/passenger.conf
/etc/apache2/mods-available/passenger.conf
/usr/lib/ruby/gems/1.9.1/gems/passenger-3.0.9/debian/passenger.conf
Each of the passenger.conf files has the following:
<IfModule mod_passenger.c>
PassengerRoot /usr
PassengerRuby /usr/bin/ruby
</IfModule>
Try to put following in apache config file
<Directory /var/www/myapp/current/public>
Allow from all
#Add following line
Options +Includes -Indexes
</Directory>
You may have mod_passenger.so symlinked into your apache2/mods-enabled directory. This would account for the double import.
Run
a2dismod mod_passenger.so
to remove any symlink, then restart Apache.
Ok
So I installed Apache httpd a while ago and have recently come back to it to try setup SSL and get it serving several different tomcat servers.
At the moment I have two completely separate Tomcat instances serving up to slightly different versions (one for dev and one for demo say) my web app to two different ports:
example.com:8081
example.com:8082
I've successfully (back in Jan) used mod_jk to get httpd to serve those same Tomcat instances to http://www.example.com:8090/dev and http://www.example.com:8090/demo (8090 cos I've got another app running on 8080 via Jetty at this stage) using the following code in httpd.conf:
LoadModule jk_module modules/mod_jk.so
JkWorkersFile conf/workers.properties
JkLogFile logs/mod_jk.log
JkLogLevel debug
<VirtualHost *:8090>
JkMount /devd* tomcatDev
JkMount /demo* tomcatDemo
</VirtualHost>
What I'm not trying to do is enable SSL.
I've added the following to httpd.conf:
Listen 443
<VirtualHost _default_:443>
JkMount /dev* tomcatDev
JkMount /demo* tomcatDemo
SSLEngine on
SSLCertificateFile "/opt/httpd/conf/localhost.crt"
SSLCertificateKeyFile "/opt/httpd/conf/keystore.key"
</VirtualHost>
But when I try to restart Apache with apachectl restart (yes after shutting down that other app I mentioned so it doesn't toy with https connections) I continuously get the error:
Invalid command 'SSLEngine', perhaps misspelled or defined by a module not included in the server configuration. httpd not running, trying to start
I've looked in the httpd/modules dir and indeed there is no mod_ssl, only mod_jk.so and httpd.exp.
I've tried using yum to install mod_ssl, it says its already installed. Indeed I can locate mod_ssl.so in /usr/lib/httpd/modules but this is NOT the path to where I've installed httpd which is /opt/httpd and in fact /usr/lib/httpd contains nothing but the modules dir.
Can anyone tell me how to install mod_ssl properly for my installed location of httpd so I can get past this error?
I found I needed to enable the SSL module in Apache (obviously prefix commands with sudo if you are not running as root):
a2enmod ssl
then restart Apache:
/etc/init.d/apache2 restart
More details of SSL in Apache for Ubuntu / Debian here.
Are any other LoadModule commands referencing modules in the /usr/lib/httpd/modules folder? If so, you should be fine just adding LoadModule ssl_module /usr/lib/httpd/modules/mod_ssl.so to your conf file.
Otherwise, you'll want to copy the mod_ssl.so file to whatever directory the other modules are being loaded from and reference it there.
Try installing mod_ssl using following command:
yum install mod_ssl
and then reload and restart your Apache server using following commands:
systemctl reload httpd.service
systemctl restart httpd.service
This should work for most of the cases.
I used:
sudo yum install mod24_ssl
and it worked in my Amazon Linux AMI.
I don't know if it is still of interest and if things have changed ever since the thread has been posted, but the /etc/apache2/apache2.conf on my system says:
Configuration files in the mods-enabled/, conf-enabled/ and sites-enabled/
directories contain particular configuration snippets which manage modules,
global configuration fragments, or virtual host configurations,
respectively.
They are activated by symlinking available configuration files from their
respective *-available/ counterparts. These should be managed by using our
helpers a2enmod/a2dismod, a2ensite/a2dissite and a2enconf/a2disconf. See
their respective man pages for detailed information.
And on my system the modules are installed in /usr/lib/apache2/modules.
I am running Ubuntu 20.04.2 LTS.