Search results

Maybe it has something to do with the VPS. bind to "*:7080" actually only bind to "209.135.157.125:7080", maybe changing the listener configuration from "*:<port>" to "209.135.157.125:<port>" can fix this problem.

By changing the owner of public/ folder. None, you need do "killall ruby", then send a request again. you should see two ruby processes, the process with larger pid is the child process, if there is only one, then the child process probably dead due to the error.

Configuration looks fine. public/ folder is owned by the same of the myrailsapp directory, right? "ps -ef | grep ruby" should show that ruby is started by the same user. Looks like you need to trace ruby process in order to find out what is wrong. do "strace -p <pid_of_child_ruby_process>"...

Please try run WEBrick under the same user of the owner of myrailsapp directory. Make sure it works properly. Check if process limit and memory limit is large enough for your application under "Rails" tab. change them to a large number.

Have you added a listener at 209.135.157.125:7080? it conflicts with the default listener "*:7080". And when you change a listener from a wildcard address to a specific IP on the same port, you need to restart it manually. After the manual restart, the "graceful restart" via web interface will...

I think it is a ruby bug or just the way it works. I cannot override a ENV[] variable in production.rb or development.rb if the same variable is set in environment.rb. Tried mongrel, same result. I tried ruby 1.8.4, 1.8.5 and rails 1.1.6. What's yours?

Is your Rails application folder owned by "root"? It is a permission problem, just change the owner of Rails application to a normal user.

Is that a fresh installation? Or upgrade? Rails app configured via the vhost template does not have anything like "RubyRailsLSAPI" or "railsLsapi". Install a fresh copy, and follow our Wiki, everything will work just fine.

Our gcc is 3.4.2, gcc command is in /opt/csw/gcc3/bin/gcc. The error is completely in the system header files. I think it is compiler environment issue.

The Wiki has been updated. :)

Are you sure? I believe values in environment.rb will override the value set in config/environments/producion.rb, etc. The same behavior under mongrel.

We don't have this problem on our solaris 10 server. What compiler are you using?

You are welcome. :) Please let me know it is fixed or not. you may need to uninstall the gem first with "gem uninstall ruby-lsapi".

Fixed in ruby LSAPI 1.7, please upgrade. http://www.litespeedtech.com/products/webserver/lsapi/ the gem file will be available soon.

It is controled by the ownership of the public/ directory of your rails application. It should be less problemetic this way. Just make sure the file ownership and permission has been set properly, you should do that no matter what, less configuration work and less chance of misconfiguration.

You can use a hard to guess URI, URI is not visible to user when LSWS perform an internal redirect, user only see the original URL to Rails, authentication can be done in Rails. There is security concern if arbitrary file location is allowed.

For that purpose, you can just set a location header with the corresponding URI to that file in the response header. Like: Location:/static/myfile LSWS will perform a internal redirect and send the file back to client, do not set status header or the HTTP status line, otherwise, LSWS will...

Looks great! Thank you very much! :)

You can try the following: Virtual Host Name: RubyVHost Domain: * Aliases: Virtual Host Root: Users/steveodom/documents/myrailsapp/trunk

Compile your own lsphp binary using PHP5 might be the fix to your problem.