I noticed that Passenger 2.0.1 became available today and I decided to upgrade a couple of my Rails sites which are using it. The process was painless but there are some configuration changes to note when upgrading from 1.0.x. So here are the steps I took.

Update: I’d also like to confirm that fair load balancing is working properly with Passenger. This is a feature that I have wanted for some time and I’m glad to see it working.

I’ve been evaluating Passenger for a while now and I’ve generally been pleased. I find that an Apache + Passenger setup has been easier to manage than alternative Rails stacks. Performance has been good and memory consumption has been lessened which also helps performance on the systems in general. However I have seen some strange bugs while using my web apps under Passenger. I am hoping that this upgrade will resolve these issues. The upgrade was relatively easy and the steps I took follows.

Just a few days ago Passenger was released in order to better integrate Rails with Apache. I decided to give it a try as I see it being very useful to simplify Rails deployments. I was also interested in testing the performance compared to nginx and mongrel.

Both of these RFC’s attempt to solve the problem that each SSL website must have its own ip. Or, in other words, they would allow name based virtual hosting for secure (ssl or tls) connections. This question sill comes up frequently as users are surprised that no one has solved this problem in the years since SSL became used for the web. The upside is that Apache now seems to support both standards, RFC 2817 in mod_ssl (2.1 and later) and RFC 3556 in mod_gnutls (0.2.0 and later). It looks like RFC 3546 is implemented in Firefox 2, Opera 8, Konqueror 4, and Internet Explorer 7. The only browser missing is Safari. So it looks like RFC 2817 is dead and RFC 3546 is the way to go. After reading up on mod_gnutls I’m excited to try it out.

• TLS for HTTP
• mod_ssl 2.2
• mod_gnutls
• SSL-enabled Name-based Apache Virtual Hosts with mod_gnutls
• SNI penetration
• Phishing for SNI progress - tantalisingly close?