Subversion 1.3.0 have already native Ruby bindings . Yahooooo!

…SWIG/ruby bindings: complete API coverage; automatic memory management; greatly expanded test suite…

Is your router properly configured?

• Download the .torrent file
• Paste the torrent link into DumpTorrent
• Copy the hash ID from its output into NatCheck page

Gentoo VPS Project – better support for VServer in Gentoo. Yahoooo!

…First, we have created the Gentoo VPS project, in order to concert the development of different implementations, and provide a convenient way for use with gentoo. The project page can be found at http://dev.croup.de/proj/gentoo-vps including Documentation and anonymous SVN access to our repository. The repository contains all patches and scripts for ebuilds in portage (openvz-sources, vserver-sources, util-vserver and vzctl), an experimental overlay, vserver stage building scripts and of course baselayout-vserver…

What’s new and interesting: support for building the vservers from stages/templates :

1. build the template from stage3 tarball

# vserver-new gentoo-template \
  --hostname gentoo \
  --context 1253 \
  --interface eth0:192.168.1.253/24 \
  stage3 /path/to/stage3-tarball.tar.bz2 <arch>
# cd /vservers/gentoo-template
# tar cjvpf /templates/gentoo-template.tar.bz2 ./

2. build vserver from the template above

# cd /vservers
# vserver-new myguest \
  --hostname myguest \
  --context 1252 \
  --interface eth0:192.168.1.252/24 \
  template /templates/gentoo-template.tar.bz2

What’s still missing: mass emerge for all vservers , so i can do vemerge—sync and vemerge system. Can this be emulated by SwitchTower task?

• SCGI Rails Runner
• SCGI and Rails on Apache 1.3.x
• Apache module named mod_scgi

And it come natively with lighttpd :)

Great tip from the Feedmarker blog :

Change the following line in your dispatch.cgi (or dispatch.fcgi, depending on which you use) file:

RailsFCGIHandler.process!

to

RailsFCGIHandler.process! nil, 1

This tells Rails to run the garbage collector and restart your app after each request (if you put the number 2 in there, it?¢‚Ǩ‚Ñ¢d be two requests, etc.). Then load your site. Now you can go back into your dispatcher file and change the setting back to normal.

Fun, and not so much. Mostly scary:

Linux in a binary world

What if.. what if the linux kernel developers tomorrow (06-Dec-2005) accept that binary modules are OK and are essential for the progress of linux.

a hypothetical doomsday scenario by Arjan van de Ven

How much we already depend from the OpenSource products?

In the latest lighttpd-1.4.8 release, they removed the debian/ support (from the changelog): removed debian/ dir from source package on request by packager

Sad :( For me i solved the problem with copying the directory from the old 1.4.7 release and fixing the changelog file. If you need better support, use bougyman’s repository. In your debian or ubuntu /etc/apt/sources.list add:

deb http://debian.bougyman.com/ unstable main

Bougyman is a very interesting person. He is working for some US goverment organization and supporting a huge Ruby On Rails system (something like a paperless office) – PostgreSQL, linux virtual server powered FastCGI listeners etc. I learned a lot from him about using RoR in production environments. Thanks. Recently I found his typo blog – All-Purpose Log for Rubyists , full with useful information. Just to mention:

• Rails-Editor gem — pure man Locomotive for Linux :) producing screen based console IDE
• spawn-fcgi in the foreground!—we are using that approach too to start/stop dispatch.fcgi listeners independently from the frontend web server (lighttpd)

via Ho John Lee?¢‚Ǩ‚Ñ¢s Weblog :

The BlackDog server turns out to be an interesting hybrid of a putting Linux on a USB flash device and putting an embedded Linux into a USB device. The actual device is around the size and weight of a pack of cards, and runs Debian Linux on a 400MHz PowerPC, drawing power from the USB interface.

Interesting to have this with Ruby and RoR installed.

In the development environment, evebody sees the debugging screen when something goes wrong. In the production environment, you only get a debugging screen, when operating from localhost.

But because of the memory leak in the current rails implementation, almost everybody now using production environment, even during the development. Action Controller have a protected method called local_request? to determine if the request is comming from a local host. Add your development machine(s) to it:

def local_request?
  [ "127.0.0.1", "192.168.1.123" ].include?(request.remote_ip)
end

An example config:

<IfModule mod_fastcgi.c>
  FastCgiIpcDir /tmp/fcgi_ipc
  FastCgiServer /path/to/myapp/dispatch.fcgi \
    -initial_env RAILS_ENV=production \
    -processes 15 -idle-timeout 60
</IfModule>

The important part is FastCgiServer . It creating static server definition – 15 fastcgi servers are started right after the server is launched and running until the server is stopped. That whay is faster that to relay on Apache to start the servers when the first time .fcgi page is accessed (dynamic server definition).

idle-timeout also is critical value. If any request takes longer than the limit, Apache will assume the FastCGI process crashed and return an error 500. If the application talks to remote servers (process requests slower) increase the idle-timeout from the default 30 to 60 seconds.