Enhancement - The control panel must always be reachable using a specific HTTPD instance

Almost finished

Re;

The nginx-panel branch is ready for testing. I would merge it but some tests are needed before. You can test it by downloading this archive: https://github.com/i-MSCP/imscp/archive/nginx-panel.zip

I'll surely merge this branch after the release of the 1.1.12 anyway. After, due to the changes, we'll surely release the version 1.2.0 directly instead of the 1.1.13.

Thanks

New Install

• Debian 6.0.9 (Squeeze): OK
• Debian 7.5 (Wheezy): OK
• Debian 8.0 (Jessie/testing): OK
• Ubuntu 10.04 LTS (Lucid Lynx): OK
• Ubuntu 12.04 LTS (Precise Pangolin): OK
• Ubuntu 12.10 (Quantal Quetzal): OK
• Ubuntu 13.04 (Raring Ringtail): OK
• Ubuntu 13.10 (Saucy Salamander): OK
• Ubuntu 14.04 LTS (Trusty Tahr): OK

Upgrade from i-MSCP 1.1.x (System without any exotic change or broken/orphan item)

• Debian 6.0.9 (Squeeze): OK
• Debian 7.5 (Wheezy): OK
• Debian 8.0 (Jessie/testing): OK
• Ubuntu 10.04 LTS (Lucid Lynx): OK
• Ubuntu 12.04 LTS (Precise Pangolin): OK
• Ubuntu 12.10 (Quantal Quetzal): OK
• Ubuntu 13.04 (Raring Ringtail): OK
• Ubuntu 13.10 (Saucy Salamander): OK
• Ubuntu 14.04 LTS (Trusty Tahr): OK

@gOOvER

Quote from gOOvER

Only one thing... if ssl is activated, there must be a redirect to https; else you get an error

Source Code
400 Bad Request
The plain HTTP request was sent to HTTPS port

No. You must simply use the correct URI scheme which is https:// . The port is not sufficient Don't worry, I've do the same mistake while my first SSL connection attempts

To resume

Normal connection ----> http://domain.tld:8080
Secure connection ----> https://domain.tld:4443

BTW: Later will be added two specific dialog boxes allowing to change those ports (of course, the default port 80 and 443 will be unallowed).

@flames

You voted "no" to the poll I don't really understand according our talk on IRC.

Re;

I've found a little bug. On new installs, php5-fpm service is not started. This is because no pool configuration file is created by default (expected since the panel is not running with php5-fpm). Thus, when a new domain is added, the reload command is failing silently and the service remains stopped. This problem occurs only in the nginx-panel branch.

In fact, the service management in i-MSCP is really bad. Indeed, the debian policy said nothing about exit codes for init scripts, meaning that the package maintainers are free to handle them as they want...Thus, we cannot rely on the exit codes to handle the services (success, failure). The same thing appear when the services are managed through upstart...

To resume

Depending of the distro, codename and package version, an init script will provide or not an exit code different than 0 when a task (start, stop...) fail. Thus, we cannot rely on exit codes to handle services.
To solve this problem, I'll provide a package which allow to handle service (start, stop, restart, reload, status...) in unified way.

Refs:

• https://wiki.debian.org/LSBInitScripts
• https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=208010

EDIT: Fixed