There is still only one solr.home instance used to load the various classes which is used as the one 'root'. >From there, you can have multiple solrconfig*.xml & schema*.xml (even absolute pathes); calling new SolrCore(name_of_core, path_to_solrconfig, path_to_schema) creates a named core that you can refer to. To refer to a named core, you call SolrCore.getCore(name_of_core) (instead of SolrCore.getCore()).
>From a servlet perspective, it seems that passing the name of the core back & forth should do the trick (so we can reacquire the correct core). One missing part is uploading a config & a schema then start a core (a dynamic creation of a core). One thing to note is that a schema needs a config to be created and it is certainly wise to use the same for schema & core creations. For the admin servlet, we'd need to implement a way to choose the core we want to observe. And the scripts probably also need to have a 'core name' passed down... I'm still building my knowledge on the subject so my simplistic view might not be accurate. Let me know if this helps. Cheers Henrib mpelzsherman wrote: > > This sounds like a great idea, and potentially very useful for my company. > > Can you explain a bit about how you would configure the various solr/home > paths, and how the different indexes would be accessed by clients? > > Thanks! > > - Michael > -- View this message in context: http://www.nabble.com/Multiple-Solr-Cores-tf3608399.html#a10084772 Sent from the Solr - User mailing list archive at Nabble.com.
