In Red5, you have pretty much full access to manipulate the embedded Tomcat engine via Spring. With that being said I would like to give details on how to change the http connector between two available options; there are several other options, but I'll only be covering NIO and BIO. First a quick explanation of these two options:

• BIO - Blocking Input / Output, this has been around since the beginning of internet time. It uses one thread per socket connnection to handle requests.
• NIO - Non-blocking Input / Output, this has been available in the JDK since 1.4. It uses a single thread to handle many socket connections.

A micro blackhole has caused a fault to form in the wall of the LHC, causing a ton of liquid Helium to spill into the tunnel. The loss of the coolant has forced a shutdown due to the over-heating of the supercooling magnets.
http://www.cnn.com/2008/TECH/science/09/20/hadron.collider.damage.ap/index.html

This is an experimental post, some of what you read in this post is fiction. Beat you to the punch Onion!

I have given up waiting for someone to claim the bounty or produce a fix for the web context logger issue. After much trouble tracking everything down, I have finally "fixed" the logging for web contexts, that is for those of you using the bootstrap method of server start up. The fixed revision is 3068 and its a little rough; meaning I still need to do bit of refactoring. I just wanted to let you all know.

Part 1

Here is what you have to put in your web.xml for each web application to get an individual logger context:

[listener]
[listener-class]org.red5.logging.ContextLoggingListener[/listener-class]
[/listener]

(more...)

I want to let everyone know about a simple RTMP load testing tool that I wrote in flex. Its really simple and allows you to rip a stream from either FMS or Red5 as quickly as possible. If you find it useful or want to add to it, let me know.

• source here
• bin-release here

The source link is for the latest version in which I have started adding shared object testing; I could use some assistance with that part if any of you have time. The bin link is to the "old" version.

I created a post about this subject almost a year ago, but there were a couple minor issues with the examples. Here I will show what eight additional months of experience can provide. The example provided here uses a custom object in Flex to pass information to and from the server, which in this case will be Red5. If one of you has an example which uses FMS on the server-side, I would be glad to include it here.
(more...)

Sorry for the delay in posting this, but I was a tad busy with the svn sync
I also had some trouble getting the IvyDE to behave as I expected. The screen cast below will show you how to get Red5 to build properly in Eclipse. One of the key points is to use Ant from a console to pull any suborn libraries down into the correct location, to do this simply turn off the auto build in your eclipse project and use this command:
ant -Divy.conf.name="java6, eclipse" dist
Note: substitute java5 for java6 if you are using JDK 1.5.
When you've finished using ant, remember to turn auto-build back on.


(more...)

I haven't heard anyone say this about Google Chrome, so maybe I'm the first. On a whim I decided to check Acid3 on Firefox 3 and it scored 71/100; then I decided to try Chrome since it rocks, and it scored 79/100. I'm really surprised that it failed, but it is still beta right?

If you'd like to try for yourself go here

Today we (Red5) moved our cvsdude repository to our existing google code repository. One of the first steps to take care of is to remove anything in your google code repository and have it reset to revision 0. To get the reset, you have to post to the google code discussion group and request it. If your repository is already empty you may see the option to reset under your projects "Source" tab.

Before you start your sync I suggest that you perform an svn update, svn commit, and svn clean up on your source repository; in that order.

Step 1: Open a terminal or command window and issue this command from top level directory of your checkout location
svnsync init --username yourusername https://myproject.googlecode.com/svn/ https://svn1.cvsdude.com/myproject

The username is for your destination which is the first url specified, the next url is for the source. You may be presented with a password prompt and / or a key accept dialog, answer and proceed

Step 2: Issue the sync command, which will be used every time you need to restart the sync
svnsync sync https://myproject.googlecode.com/svn

The url should point to your destination.

If everything goes smoothly, you will not see any errors and your repository will exist on the destination. This was not the case for us, so here are the ways we got past the issues that resulted.
(more...)