Tim, New documents are automatically distributed across the available forests. Document placement doesn't pay any attention to forest size: the algorithm uses the number of available forests and the URI of the document. If you want to see which forest will be selected for a new document, you can use the new 4.2 function xdmp:document-assign: http://developer.marklogic.com/pubs/4.2/apidocs/UpdateBuiltins.html#xdmp:document-assign
If you have existing forests that aren't well-balanced, you can use XQSync or Corb to re-balance them. -- Mike On 24 Jan 2011, at 06:33 , Tim Meagher wrote: > Hi Folks, > > When a database is spread across multiple forests, by what means does the > MarkLogic server attempt to distribute content – does it try to split the > content evenly across the forests or does it attempt to evenly distribute the > content based on the amount of available memory in each forest? For example, > if database DB1 consists of 2 forests, F1 and F2 where F1 is on disk 1 and > has 100G of free space, but F2 is on disk 2 and has 10G of free space, and I > add 10G of new content to DB1, what should I expect to happen if I am not > designating the destination forest when inserting the content? Will all (or > most of the content) go to F1 or will it be evenly distributed between F1 and > F2? > > Thank you! > > Tim Meagher > > _______________________________________________ > General mailing list > [email protected] > http://developer.marklogic.com/mailman/listinfo/general _______________________________________________ General mailing list [email protected] http://developer.marklogic.com/mailman/listinfo/general
