On 03/12/2017 06:21 AM, Srini Seetharaman wrote: > Hi, I have a setup where all instances (running stable/Boron) of the > cluster expose a particular RPC implementation. App1 registers as the > provider in its init() and App2 gets a handle of the RPC service in its > init(). Thus my hope is to just keep it as a local RPC call. > > However, just because instance A comes up a few secs before instance B, > invocation of a RPC call at B is forwarded to the instance A. This > throws off my logic of keeping the invocation local. Is there a way to > force the RPC call to be local, even if it meant throwing an exception > about no implementation being available yet?
RPC clients should not care about provider location (same as a provider should not care about client location), as that introduces strong coupling between them. In case of multiple providers being present, routing should occur to the 'closest' registered provider, which is something we fixed as part of BUG-3128 and will be part of Boron-SR3. Can you test against a current snapshot? Regards, Robert
signature.asc
Description: OpenPGP digital signature
_______________________________________________ controller-dev mailing list controller-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/controller-dev
