On Tue, Jan 24, 2012 at 7:33 PM, Justin Deoliveira <jdeol...@opengeo.org> wrote: > > > On Tue, Jan 24, 2012 at 3:26 PM, John Armstrong <john.armstr...@ccri.com> > wrote: >> >> On Tue, 24 Jan 2012 15:21:15 -0700, Justin Deoliveira >> <jdeol...@opengeo.org> wrote: >> > And sorry, by disable i mean remove the jars. Or perhaps just the >> diskquota >> > jar might do it, not 100% sure. >> > >> > On Tue, Jan 24, 2012 at 3:20 PM, Justin Deoliveira >> > <jdeol...@opengeo.org>wrote: >> > >> >> Hi John, >> >> >> >> Unfortunately the dbconfig module is sort of unmaintained at the >> moment. >> >> That said, it looks like gwc is probably doing some initialization in a >> >> separate thread, which the hibernate startup does not like. >> >> >> >> If you don't need geowebcache, i would just disable it for now. >> >> >> >> -Justin >> >> Thanks; just to be sure I understand you, you think that it's something in >> GWC that's conflicting with the Hibernate module? I'm not sure I >> understand where you're getting that, but I'm willing to believe it. > > > Sorry, allow me to explain better. The way hibernate is set up in geoserver > is that a session is bound to a single request, or thread. If you try to > access a hibernate backed bean that has a lazy property from outside that > thread/session then you get this exception. > > Now, during regular geoserver operation this is rarely an issue because > essentially there is a filter in place that always set ups and tears down > the thread local session, meaning that for the life of a single > request/thread a valid session will always be present. > > However, on startup we don't have this same filter available, so we fake it. > BAsically the code that loads the configuration creates a session before > loading, and tears it down after loading. Again this session is bound to the > current thread, which means if any components that initialize on startup do > so in a separate thread, and they access a lazy property of a bean, then > this exception occurs. > > So yes, i am saying that gwc is doing just that, it is using a new thread to > do some of its initialization on startup. Gabriel should be able to verify > this but i am pretty sure i have run up against this one before. > > Hope that makes it a bit clearer.
Right. My understanding is that anything that calls the catalog outside an http client request will suffer from this. GWC is an example, as it may also do that if you want to seed the cache. Another example is could think of, but didn't really tried, is the wps module, as it can run async processes. I don't know what the general practice to avoid that kind of lazy initialization errors with hibernate is though. Cheers, Gabriel > > -Justin > > -- > Justin Deoliveira > OpenGeo - http://opengeo.org > Enterprise support for open source geospatial. > -- Gabriel Roldan OpenGeo - http://opengeo.org Expert service straight from the developers. ------------------------------------------------------------------------------ Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Geoserver-devel mailing list Geoserver-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/geoserver-devel
