Hi again, I did my first tests with spread, and it seem to work. But there is one issue with the properties I think.
I do need for every group an own hedera.channel.spread.private_name, am I right? In my eyes I would need a bunch of property files, for every virtualdatabase an own, right? Is there a way to configure this by the vdb.xml? Because I try to automate the process of creating our virtualdatabases and I would prefer having one spread.property file for all and just adapting the vdb.xml. Do you understand my problems? Thx for your help! Regards, Malte > Hi, > > On May 28, 2007, at 3:48 , Malte Altmann wrote: > >> Hi Nuno, >> >> I'll try it tomorrow, sounds very simple, but we'll see ;) >> > > ok. > >> Are there advantages concerning the memory usage or other >> performance issues? >> > > maybe it will use less memory because Spread uses a different > approach: the java side of the code is a very simple client that just > connects to the daemon to send/receive messages. The high tech is all > in the daemon, not in the client that is used by a java spread > application (in this case, the controller). So, off course the > controller process will spend less memory, part of the code is being > executed in another process. > > I think that you will not have a great performance because of their > Java API. Something is not good about this API and makes it run very > slow. We after built another java API for Spread and runs quite > faster. See http://jgcs.sf.net. This is a generic API also with > several implementations and one of the bindings is for Spread. To > implement that, we used JNI. > > regards, > nuno > > >> Best regards, >> >> Malte >> >> Nuno Carvalho schrieb: >>> Hi, >>> >>> the reasons for this implementation were simple: I needed it, so I >>> implemented it. I had the code, so I just decided to contribute it >>> to the forge :) >>> >>> just install the spread daemon on the machines were you will run >>> the controller, get hedera from CVS, compile it and add the jars >>> to Sequoia. use the hedera_spread.properties file instead of the >>> hedera_jgroups.properties or hedera_appia.properties. >>> >>> Cheers, >>> Nuno Carvalho >>> >>> On May 27, 2007, at 7:34 , Malte Altmann wrote: >>> >>>> Hi @all, >>>> >>>> today I (I think the first time) read something about an Spread >>>> implementation of Hedera. >>>> What are the reasons for this implementation, are there benefits >>>> compared to using appia? >>>> How can I easily test a bit with spread? Is it possible to change >>>> to spread, as easy as it was to appia? >>>> >>>> Best Regards, >>>> >>>> Malte >>>> _______________________________________________ >>>> Sequoia mailing list >>>> [email protected] >>>> https://forge.continuent.org/mailman/listinfo/sequoia >>> >>> >>> -- >>> Nuno Carvalho >>> University of Lisbon, Portugal >>> http://dialnp.di.fc.ul.pt >>> >>> >>> >>> _______________________________________________ >>> Sequoia mailing list >>> [email protected] >>> https://forge.continuent.org/mailman/listinfo/sequoia >>> >> _______________________________________________ >> Sequoia mailing list >> [email protected] >> https://forge.continuent.org/mailman/listinfo/sequoia > > > -- > Nuno Carvalho > University of Lisbon, Portugal > http://dialnp.di.fc.ul.pt > > > > _______________________________________________ > Sequoia mailing list > [email protected] > https://forge.continuent.org/mailman/listinfo/sequoia > _______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia
