StormGrind | CirrAS | Setup Testing

You have downloaded (or build yourself) CirrAS appliances but now you're thinking  how to check if everything works fine?

node-info

We prepared a small web application called node-info that prints useful informations about client and server (IP addresses,  ports, selected JBoss AS profile). You can build it yourself or download  it from our server.

Building node-info from source

Web-app source code is located in CirrAS sources in docs/node-info/ directory. Just  grab it and execute

mvn clean package

in node-info/ directory. It will produce a  node-info.war file in node-info/target/ directory.

Downloading node-info from our server

Latest node-info build is located here.

Installing node-info

Installing node-info application is simple, just copy it to JBoss AS deploy folder for selected profile: $JBOSS_HOME/server/$PROFILE/deploy/.  For help on selecting right profile take a look at StormGrind | CirrAS | JBoss AS Profiles.

Using mod_cluster manager

You can display all nodes in cluster using mod_cluster manager. It is  a great, simple web-app that shows many useful informations about you  cluster. It is by default enabled in new CirrAS versions.

 

 

 

Note: mod_cluster manager is enabled by default in all CirrAS Beta versions. After CirrAS becomes GA it'll be disabled by default.

Step-by-step troubleshooting

First install node-info application as described above.

 

An example is best to describe the process. Let's assume you have two JBoss AS 5 nodes (back-end-appliance) and one load balancer (front-end-appliance).  IP addresses:

  • front-end-appliance: 172.16.10.131
  • back-end-appliance 1: 172.16.10.132
  • back-end-appliance 2: 172.16.10.136

 

First of all check if every JBoss AS in cluster is up; point your  browser to each JBoss AS node on 8080 port (http://172.16.10.132:8080 and http://172.16.10.136:8080 in my  case). If there is a blank page – everything is OK (as ROOT application was removed from nodes). If not, check network preferences in appliance.

 

Next step is to check if load balancer is working. Point your browser  to your front-end-appliance address (http://172.16.10.131 in my case). There should be waiting for you shinny Apache test page.

 

If everything is working, you deserved a beer, but don't drink too much, because we have still a few steps.

 

Install on all nodes node-info application as  described in above section. After that point your browser to /node-info context on all nodes (http://172.16.10.132:8080/node-info/ and http://172.16.10.136:8080/node-info/ in my case) it should display information about current node an client  browser.

 

You can list all nodes in cluster using mod_cluster manager. See  above.

 

Last step is to check if balancing is working. Point your browser  again to load balancer but with /node-info context (http://172.16.10.131/node-info/).

 

 

It should display node information. Take a closer look at displayed  addresses.

 

Now you can stop node that you're connected to (172.16.10.132 in my case). After you stopped the node (powered off, killed JBoss, etc) refresh your browser, it should display another node information,  like this:

 

 

Now, you can deploy your apps and be a proud CirrAS user, but if  you have any troubles with getting it to work, contact our community. You can also create  a new ticket.