Adam Lassek's questions

I'm attempting to deploy a Rails 3.1 application to an Ubuntu 11.04 server running nginx-1.0.0 and passenger-3.0.7

I have a system-level RVM install at /usr/local/rvm/gems/ruby-1.9.2-p180

My nginx.conf looks like:

http {
    passenger_root /usr/local/rvm/gems/ruby-1.9.2-p180/gems/passenger-3.0.7;
    passenger_ruby /usr/local/rvm/wrappers/ruby-1.9.2-p180/ruby;

    # ...

    server {
        listen 80;
        server_name my-domain.com;
        root /srv/www/my_project/current/public;
        passenger_enabled on;
        rails_env production;
    }
}

The problem is that I have two gems in my Gemfile that point to git repos. Bundler says that they are installed, they are present in RVM's gem directory, but Passenger throws an error:

[email protected]:my-gem (at master) is not checked out. Please run `bundle install` (Bundler::GitError)

Running rails console from the deployment directory works fine. What am I missing?

I'm designing a JSON API, and I'd like to version the API using content negotiation of some kind. I'm currently planning on using Vendor MIME Types to do this.

While I can definitely do this at the application level, I'm thinking it would be best to make this happen at the HTTP server level. Is this possible with Apache or nginx?

The Content-Type would look something like: application/vnd.vendorname-v1+json or possibly using parameters: application/vnd.vendorname+json;v=1

I have an Ext3 filesystem sitting on an LVM volume group. The partition was formatted before online resizing was enabled by default.

Is it possible to turn this feature on after the fact, like journaling, or does it have to be present from the beginning?

I'm attempting to remove an old 300GB drive from my volume group and replace it with a shiny, new 1TB drive.

The pvmove operation ran until 99.9% and then aborted. I believe this is because I ran pvcreate on /dev/sdf directly instead of creating a partition. My problem is, it moved the physical extents to /dev/sdf but the original drive, /dev/sdc1 is still in the volume group, and they both have identical extent counts. It seems that my vg has a duplicate set of extents.

Every guide for removing a disk from a vg involves a pvmove and a vgreduce. Since the extents are duplicate, I don't want to move them; I need to remove them from the vg entirely. How should I do this without damaging the integrity of the existing volume group?

mstsc allows you to save your credentials for a system so you don't have to type them in every times, which is nice when you use randomly-generated long passwords.

My Server 2008 machine allows this just fine, but on my Server 2003 system it fails, with the message:

The system could not log you on. Make sure your User name and domain are correct, then type your password again.

I'm positive the credentials I'm saving are correct, and when I type it in by hand it works.

I have spent the last several hours attempting to get Ruby on Rails running on a Windows server with no luck.

At first I tried configuring a test application through IIS7's FastCGI support, but the documentation for this is not very good. I've been following this blog entry, and this one, and this one, and this one but everything seems to be missing major steps, or are out of date. And every article keeps linking back to this Howto from rubyonrails.org that doesn't exist.

The sense that I'm getting is that even if I manage to make this work, IIS' FastCGI isn't good enough to use in a production environment anyway.

So it looks like my best bet is to setup a reverse proxy in IIS that points to Apache & Mongrel/Passenger using ARR and UrlRewrite. Is there anybody else out there stuck deploying a Rails application on a Windows stack? Am I on the right track? Can you give me a better idea of how to configure this?

I believe Plesk already installed an instance of Apache/Tomcat running on this server using a different port, so adding another virtual host shouldn't be difficult; the hardest part seems to be setting up the reverse proxy through IIS.

Addendum

Since posting this, I have gone on to build a virtual server running Linux and serving our Rails products with Apache, Passenger and Ruby Enterprise Edition.

Linux is so easy, cheap and reliable an option that trying to jam the square peg of Rails into the round hole of Windows/IIS is a fool's errand.

The only possible exception to this is if your Windows env. supports Java, and you can use JRuby. By all accounts it is one of the better Ruby implementations. Otherwise, give up and save your sanity.

I'm very new to managing Cisco equipment, so bear with me.

I'm configuring a Cisco ASA 5505 router for my office, and I am reasonably competent enough with the console to configure the basics -- our business needs are not extravagant. Our pings are being dropped by the router, however.

How do I configure the router to allow ICMP Echo Requests? Are there other types of ICMP requests that should be allowed? What are the potential downsides of allowing them all?