JB AS running correctly is a grey zone. There are many components and cyclical inter-dependencies, it’s hard to know if they’re ever configured correctly without access to middleware developers. Basically, JB can break in 100^100 ways and still appear to work.

Packages would be better, but then my company wouldn’t be able to charge $$$ for installing jboss. ]]> http://candrews.integralblue.com/2009/07/install-jboss-4-2-on-centosrhel-5/comment-page-1/#comment-54 LM Wed, 05 Aug 2009 09:55:08 +0000 http://candrews.integralblue.com/?p=88#comment-54 Very good work. I've learned much more in one hour reading and installing Jboss this way than surfing hours on jboss wiki/forum. Many Thanks, Craig ! Very good work. I’ve learned much more in one hour reading and installing Jboss this way than surfing hours on jboss wiki/forum.
Many Thanks, Craig ! ]]> http://candrews.integralblue.com/2009/07/install-jboss-4-2-on-centosrhel-5/comment-page-1/#comment-51 Karsten Wade Wed, 01 Jul 2009 21:22:20 +0000 http://candrews.integralblue.com/?p=88#comment-51 Excellent post and very well argued. Let's hypothesize for a moment. Let's say that JBoss engineers are i) busy working on the next release or something, ii) are not hearing from their customers in overwhelming numbers that packaging matters, and iii) maybe that is partially due to having an ear tuned mostly to developers who tend to care less about the advantages of package management. (BTW, the above situation describes the early Fedora Core days when it was time to work on a RHEL release. It was from this situation that Fedora Extras was born, and that changed the game forever. What is the equivalent 'JBoss Extras' here?) What if people in a similar situation as yours were to speak up and say, "Here, we'll do that for you." Granted, it's not going to be easy, nor necessarily possible with all JBoss projects/products. The binary JAR compiled with another binary JAR is a bit of a rabbit warren to chase through. Changes may be required in core projects that are hard to bring about. But there are various grassroots efforts, home-baked packages, and so on where other people are working on this. We've had some push on it from the Fedora ISV special interest group[1]. From a community angle, I see the work is two-fold: 1. Some person or group needs to lead the effort, from a project management and technical stand point. Doing this in Fedora makes the most sense; work there can flow in to EPEL[2], and sometimes then in to RHEL directly. This does not need to be an @redhat.com/@jboss.com person. 2. People to work the problem from various angles. There is going to be a bit of low-hanging fruit, some medium-hard, and an amount of very hard nuts to crack. Some of these people must be @redhat.com/@jboss.com, but maybe not as many as we think/fear. What if ... all of that? What next? [1] http://fedoraproject.org/wiki/ISV_Special_Interest_Group [2] http://fedoraproject.org/wiki/EPEL Excellent post and very well argued.

Let’s hypothesize for a moment. Let’s say that JBoss engineers are i) busy working on the next release or something, ii) are not hearing from their customers in overwhelming numbers that packaging matters, and iii) maybe that is partially due to having an ear tuned mostly to developers who tend to care less about the advantages of package management.

(BTW, the above situation describes the early Fedora Core days when it was time to work on a RHEL release. It was from this situation that Fedora Extras was born, and that changed the game forever. What is the equivalent ‘JBoss Extras’ here?)

What if people in a similar situation as yours were to speak up and say, “Here, we’ll do that for you.”

Granted, it’s not going to be easy, nor necessarily possible with all JBoss projects/products. The binary JAR compiled with another binary JAR is a bit of a rabbit warren to chase through. Changes may be required in core projects that are hard to bring about. But there are various grassroots efforts, home-baked packages, and so on where other people are working on this. We’ve had some push on it from the Fedora ISV special interest group[1].

From a community angle, I see the work is two-fold:

1. Some person or group needs to lead the effort, from a project management and technical stand point. Doing this in Fedora makes the most sense; work there can flow in to EPEL[2], and sometimes then in to RHEL directly. This does not need to be an @redhat.com/@jboss.com person.

2. People to work the problem from various angles. There is going to be a bit of low-hanging fruit, some medium-hard, and an amount of very hard nuts to crack. Some of these people must be @redhat.com/@jboss.com, but maybe not as many as we think/fear.

What if … all of that? What next?

[1] http://fedoraproject.org/wiki/ISV_Special_Interest_Group

[2] http://fedoraproject.org/wiki/EPEL ]]>