Heads Up :
The post you are writing will appear in a public forum. Please ensure all content is appropriate for public consumption. Review the employee guidelines for the community here.
Created by: null on 04-12-2003 03:38:10 PM This is being posted from another source.
What is the best manner to setup (i.e. development vs. run-time) lab and production environments? We would like to deploy new versions of code without necessarily having a development environment installed into the production environment. We want to retain control over where the code gets developed and where it gets deployed.
Subject: RE: Development vs. runtime deployment Replied by: null on 04-12-2003 04:10:02 PM Hi,
This is a good question and Audium customers have taken various approaches to address this question. When you make a new installation of Audium, it installs both the Audium Server (runtime component) and the Audium Builder (development component). In lab environments, this is fine as the developer will need to spend a lot of time with the Builder and will need to test their output by using the Server locally. Additionally, development tools would be installed as well, such as tools for editing and compiling Java. Typically, we see lab environments loaded with software the developer needs without a lot of concern for call performance as there usually aren't more than a couple of phone calls being handled by the system.
In production environments, the desire is to load only the software that is essential for the system to function. The more bare-bones the better since you maximize the performance and minimize the variables that can cause conflicts or systems issues. To that end, one can make a new installation of Audium on a production machine and delete those directories in the Audium Home directory that are not directly related to the runtime component, for example the builder directory that houses the Audium Builder. While there are no performance hits at all associated with having the Builder installed on a production machine, this would be done as a way of limiting development work (even if it is as simple as changing the name of an audio file) to the lab, not production. Even if it is possible, you don't want to be able to make changes to an application on the production machine. Here is a list of the directories it is safe to delete without affecting the Audium Server:
agent (only if you don't plan on using SNMP. If you do, this folder must remain).
builder
docs
UninstallerData
The rest of the directories hold files necessary for the running of the Audium Server (along with the Audium.war file installed on top of a supported Java application server).
You also mentioned deployment. Once you have set up the development and production environments as you prefer, it is possible to automate the process of deploying from the development environment to the production environment. Everything that is necessary to define an Audium voice application are:
The folder containing that application in the Audium Home applications folder. This folder contains the call flow, settings, XML decisions, and application-specific Java classes for the application.
The common folder. Any Java classes that are made available to all applications on the server are deployed here so if the application uses any of these classes, the contents of the common folder need to exist in both development and production environments.
Any external files required by the application. These include audio files, grammar files, and any miscellaneous files loaded by the application must also exist in both environments. Typically, these files are static and for best performance, should be hosted from a separate web server (as opposed to an application server). This is because application servers are designed to handle the generation of dynamic content and are less efficient than standalone web servers for serving static content.
So with this in mind, one can simplify the process of deploying by making a script, for example, that zips up all this content, transfers the archive to the production environment, then unzips it and distributes the files to appropriate locations. In fact, even doing this manually requires very little time. As long as this becomes a set procedure, one can make all changes to an application in the development environment and deploy to the production environment in this way. Giving the process structure will help you to control it.
Audium is working on features for future versions of the software that will aid in the deployment of an Audium voice application from one environment to another in a more automated fashion. Any suggestions on this would welcome.
Hope this helps!
Subject: Development vs. runtime deployment Replied by: null on 16-09-2004 07:22:58 AM Hi,
Is the only/recommended way to setup a runtime-only environment to perform a full install and then delete those specific directories?
Is this what's meant by the term "headless"?
Thanks in advance.
Regards, Strafford Walton
Subject: RE: Development vs. runtime deployment Replied by: null on 17-09-2004 02:42:25 AM As of version 3.3, this is the only way since the installer only installed both Server and Builder.
In 3.4, the installer has been improved to give the user a choice of whether to install Call Services (formerly called the Audium Server), Studio (formerly called the Audium Builder), or both. An installation involving just Call Services would be called "headless", because Studio, being the interface by which voice applications are constructed - i.e. the "head" of the software, is not installed. You would only be able to run pre-built applications on a headless system and would be unable to make changes on it. This is why typically it is QA and production setups that are headless because one is not supposed to make changes to applications directly in these environments.
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: