On Mon, Dec 7, 2009 at 6:01 AM, ant elder <[email protected]> wrote: > On Sat, Dec 5, 2009 at 1:00 AM, Raymond Feng <[email protected]> wrote: >> Hi, >> >> I fixed the issue under http://svn.apache.org/viewvc?rev=887470&view=rev. >> > > Thanks that does help get passed this particular problem. > >> I also have some local changes that enable the build.xml to run two nodes >> with different JVMs over tribes based multicast. Do you know if the build >> machine supports multicast? If so, I can commit the changes. >> > > I'd guess that multicast will work within the one build machine, but > if it didn't the static routes support we have does work enough to > support this as there's just two nodes. However, the test isn't really > good enough at verifying things work yet - if something breaks the > test will still pass successfully - so there's not a whole lot of > point having it part of the build until we fix that. I've already > started looking at adding some better tests as part of all this domain > work. > > Incidentally i'm finding quite a lot of other issues as i use this to > do with things like the classloaders using by tribes or sca bindings, > inconsistencies in the endpoint URIs used by all the different pieces, > particular runtime requirements caused by the way endpoints get > serialized etc. I'll try to post more on those in separate emails. > > ...ant >
Where we are with this issue ? Does this happen just when deployed in the tomcat/tuscany.war environment or is there a way to reproduce this using some itest (e.g. itest-nodes-two-nodes-two-vms-test) ? If you guys still have the issue and want some help, I could take a look at this. -- Luciano Resende http://people.apache.org/~lresende http://lresende.blogspot.com/
