Most of the conversation related to DB-oriented persistence is a good option, but the file-based approach does works quite well and is pretty stable. In addition, if you are using the lucene indexing on your fields, the search performance is definitely good and probably on par with DB (I haven't compared the two yet). One less dependency/software-system to cause issues is a good thing ;-)
As for a previous remark regarding a lot of nodes in a single hierarchy/flat-storage, that is a very common problem and also exists in Slide -- Jackrabbit is not unique to this problem. Jackrabbit is definitely very good and maturing very well. My only open complaint is still waiting for full WebDAV support with DASL <basicsearch> support -- I tried it myself, but it ended up being a horrible hack that I threw away ;-) http://issues.apache.org/jira/browse/JCR-488 > He's referring to the fact that properties are stored in > separate records. Thus, if you save a node with 10 > properties, you end up with > 11 database records. CRX and some other proprietary > persistence managers save the whole node in a single record, > achieving a major performance boost for some use cases. >
