: As it pertains to Solr, I've often used Master and Searcher. : Probably even more correct would be Indexer and Searcher. : Primary and Secondary don't quite sound right for the Solr : situation... (but Master and Slave doesn't capture it any better : either).
primary/secondary doesn't relaly apply because the labels are too vague ... primary for what? primary search box? primary indexing box? typically the terms primary/secondary relate to failover modes, you have a primary woozle that does everything a woozle is suppose to do, but if htat woozle stops working the secondary woozle steps in and acts very woozly. This is not an inherient concept in Solr. the (computer) master/slave concepts are not inherient in Solr either -- but they are a part of the *distribution* of indexes that Solr has hooks for. one Solr instance can be declared the "master" of the index and given all the updates to process while other Solr instances can choose to "slave" off of a master of their choice and take indexes as is from that master -- but even then the "slaves" may themselves be "masters" as far as other solr indexes further down a distribution chain are concerned -- so even using hte terminology "master/searcher" doesn't really apply, since the "slaves" may not actaully be used for seraching, but only as way-points. In the end, it's all fairly irrelevent. While the term "master" is used quite a bit in the distribution scripts (to indicate where to pull an index from) the term "slave" is the only contentious term in the pair and it's use is confined to in distribution.jsp and a few comments in the scripts about snappuller.status. if anyone wants to submit a patch that changes the comments/variables names, feel free -- just as long as it's something descriptive regarding the nature of the data exchange relationship. -Hoss
