Hi, right now, they are temporarily cached, but the entries are stored on the disk.
We have rewrote the API to manage binary and string values, with some extension points for streamed values (both binary and string), but this has not been implemented yet. Now, depending on what you do, you can still manage big values. It's just a matter of limitating the number of entries you cache (this is configurable), and to avoid indexing those attributes. You also have to augment the JVM dedicated memory. Hope it helps. On Wed, Jan 14, 2009 at 1:18 PM, Chuan Li <[email protected]> wrote: > Hi guys, > I have the assignment to save/read large character attribute ( ~ 10 MB) on > your server. The system has about about hundred users at the same moment. > This question is that, do you think it is practical to do so? How does the > server deal with large character attributes, are they loaded in the memory > first or simply streamed? > > Thanks for your kind answer! > > Chuan > -- Regards, Cordialement, Emmanuel Lécharny www.iktek.com
