Experience for management of tons of xml doc using exist-db

On Wed, Jan 15, 2014 at 1:51 AM, easy <lin_xd at 126.com> wrote: > Hi,all, > Based on my discontinuously near ten years experience using exist-db, I > made a plan to using exist-db to management more than millions people's EHR, > till now, it almost can work, here is what I have done: > 1) split the large quantity of doc into small collection( not one > collection for one people, but one doc for one people), target to speed > query personal info and insert new record nodes. > 2) with the kind developer 's help , we can using new range index to quick > query big data based on some conditions. > 3) using the backup policy to do backup. > ... > But, there are still some difficult to worry about: > 1) exist-db often caused corrupt for a heavy load query, and this often > cause db recover fail. and how to avoid heavy load query to affect other > query? > 2) when will use re-index? after corrupt? the re-index process is very > time consumed. how to accelerate re-index? > 3) how to control concurrently update db? often this cause the db stop > responding to xquery, and as 1). everyday there will more than 1000000 > files add to the db, so this is a difficult problem. > 4) how to control backup copys ? the backup file will cost large disk > space if all old copy exist. > > Thanks for advice > > > > > -- > 此致 > > easy > > 莫愁前路无知己，天下谁人不识君。 > > > > _______________________________________________ > BaseX-Talk mailing list > BaseX-Talk at mailman.uni-konstanz.de > https://mailman.uni-konstanz.de/mailman/listinfo/basex-talk >