Thanks a lot Stefan. So, the main idea is to preserve the package and to configure the schema parameter.
./alex -- .w( the_mindstorm )p. On 4/22/06, Stefan Guggisberg <[EMAIL PROTECTED]> wrote: > On 4/22/06, Alexandru Popescu <[EMAIL PROTECTED]> wrote: > > Hi! > > > > I am wondering how this customization of the DDL should be done... I > > must go in the jar and modify the DDL in there? Or is it a way to > > externalize it? > > you could e.g. > > - copy the derby.ddl, modify it and save it as e.g. derby_custom.ddl > - put this .ddl in a jar file (package o.a.j.c.state.db) and add the > this jar file > to your classpath > - make sure you configure your PersistenceManager to use the custom > schema (i.e. <param name="schema" value="derby_custom"/> > > see > http://jackrabbit.apache.org/apidocs/org/apache/jackrabbit/core/state/db/package-summary.html#package_description > http://jackrabbit.apache.org/apidocs/org/apache/jackrabbit/core/state/db/DerbyPersistenceManager.html > > cheers > stefan > > > > many thanks in advance, > > > > ./alex > > -- > > .w( the_mindstorm )p. > > > > > > On 4/20/06, Alexandru Popescu <[EMAIL PROTECTED]> wrote: > > > Hi Stefan and thanks for the explanations! > > > > > > There is no panic, but everybody should be able to understand this before > > > going into production :-). > > > > > > I guess this is very good knowledge for everybody: > > > > > > [quote] > > > 30k child nodes to one single node per normal blob entry. If you think > > > that your structure will > > > overpass this number (which is completely unrecommended) you can > > > customize the blob size (at least > > > with MySQL, Derby) > > > [/quote] > > > > > > > > > ./alex > > > -- > > > .w( the_mindstorm )p. > > > > > > > > > #: Stefan Guggisberg changed the world a bit at a time by saying (astral > > > date: 4/20/2006 12:02 PM) :# > > > > don't panic, folks! ;) > > > > > > > > the default schema for derby uses a 'blob' column for storing the > > > > serialized node states. as brian pointed out correctly, the blob > > > > column has a max. size of 1mb if no size has been specified. > > > > > > > > i ran a quick test: i was able to add ~37k child nodes before hitting > > > > the 1mb size limit. i think that this is perfectly reasonable and > > > > i guess that most repositories won't ever come near this limit. > > > > > > > > however, if you really think that you need to be able to add more > > > > than ~30k child nodes to one single node you could easily change > > > > your schema, e.g. by specifying 'blob(5M)' as the data type of > > > > the NODE_DATA column.. > > > > > > > > again: i would stronlgy discourage from this kind of 'flat' repository > > > > model unless you have a very compelling reason. jcr's data model is > > > > inherently hierarchical so i'd suggest to better make use of this > > > > powerfull > > > > feature. > > > > > > > > cheers > > > > stefan > > > > > > > > > > > > On 4/19/06, Brian Moseley <[EMAIL PROTECTED]> wrote: > > > >> On 4/19/06, Martin Perez <[EMAIL PROTECTED]> wrote: > > > >> > > > >> > I have no problems uploading various megabyte files to derby without > > > >> > configuring anything... > > > >> > > > >> that's not the issue. saving serialized node states with > > > >> DerbyPersistenceManager is what we're talking about. > > > >> > > > > > > > > > > > > >
