Red Hat Bugzilla – Bug 1017239
[Doc Bug Fix] mod_cluster and clustering configuration mixed up together
Last modified: 2015-10-27 16:20:12 EDT
if you take a look at Procedure 16.7. Configure a Worker Node,
you might notice that it seamlessly blends to clustering configuration.
It is not wrong per se, yes, one might carry out the steps mentioned, yet, it has little to do with mod_cluster as such.
It is actually a clustering configuration, technically absolutely unrelated to mod_cluster. As soon as you configure mod_cluster subsystem, it's ready. No need for any clustering configuration stuff.
It must be clear to the customer that:
1) As soon as you configure the mod_cluster subsystem in Jboss EAP 6 instance and mod_cluster module on Apache HTTP Server instance, you are ready to go with load balancing. No matter whether you are using standalone not clustered servers, clustered standalone servers, not clustered domain managed hosts, clustered domain managed hosts...
2) If you want to have session data replicated among cluster nodes, you must configure clustering subsystem. The default of standalone-ha, standalone-full-ha, or in domain, ha or full-ha profiles is good enough for start (provided you have <distributable/> element within your web.xml in your webapp you want to have replicated).
TL;DR: mod_cluster subsystem and clustering subsystem configurations are _absolutely_ unrelated and _not depending_ on each other. Luring the customer into thinking otherwise is confusing.
Dear Doc guys, could you look into it please?
Dear Tom, regarding my comment 1, may I ask you to check who's working on this bugzilla and whether there has been any change in the text of our current documentation build?
With more and more changes piling up, it's gonna be harder and harder to locate what was supposed to be fixed...
In JIRA -> https://issues.jboss.org/browse/JBEAP-1300