On May 16, 2013, at 12:00 PM, Mircea Markus <mmarkus(a)redhat.com> wrote:
Hi Galder,
Whilst reviewing Tristan's pull request for ISPN-3008[1] I saw that we allow
configuring "fetchInMemoryState" for topology caches and wondering why we do
that?
^ Not sure exactly what you mean… When a node joins in, it needs to receive the topology
from other nodes to be able to provide topology headers with the correct info to the
clients. That's why the topology caches can be configured with fetching memory.
Alternatively, since the cache is replicated, a similar effect can be achieved with a
cluster cache loader. This was introduced a while back as a result of a bug with state
transfer (I think it was something related to
https://issues.jboss.org/browse/EDG-44)
Shouldn't it be enabled by default/enforced?
^ Either that, or the cluster cache loader are used, both of which serve the same
purpouse.
--
Galder Zamarreño
galder(a)redhat.com
twitter.com/galderz
Project Lead, Escalante
http://escalante.io
Engineer, Infinispan
http://infinispan.org