On 4/8/11 8:53 AM, Heiko W.Rupp wrote:
Am 08.04.2011 um 15:48 schrieb Brian Stansberry:
> This is what the runtime-name is for. I'm trying to make the hash
> disappear as something the end-user cares about, e.g. the underlying
> /deployment=foo.war:add operation can itself accept the content and
> upload it, thus removing the need for the end user to care about the hash.
I don't really care about the hash here.
If I upload test.war v1 and v2, both will be named at upload test.war and
as runtime name test.war, as they are supposed to go into different server groups.
Previously they would have been on two (physically) different clusters, but now
one can expect that users will want to set up one server group for test and another
for integration where they more or less deploy the same stuff into.
So you upload them as test.war.v1 and test.war.v2 both with runtime-name
test.war. They are two distinct deployments that need to be managed
separately, and their name is their unique handle in the management
system. The runtime-name is how the non-management layers in the running
server (i.e. the deployment processors and subsystem services) see the
deployment.
--
Brian Stansberry
Principal Software Engineer
JBoss by Red Hat