On Wednesday 02 December 2015 10:00:51 Chetan Mehrotra wrote:
> On Tue, Dec 1, 2015 at 10:07 PM, H K <[email protected]> wrote:
> > My problem is that when I write data to the repository through my
> > application, Sling does not seem to get it fast enough, which may be
> > because it is caching or because it only periodically checks to see if
> > the data it has is up-to-date or not

hi Chetan,

> What you are observing here is the result of eventual consistency
> model of Oak which is a MVCC based store [1]. When using
> DocumentNodeStore (used for Mongo) each cluster node periodically
> polls the backend for change in root node and then updates its head
> revision to make those changes visible to current cluster node. So
> change done by one cluster not would not be immediately *visible* on
> other cluster nodes.

but this is not the case for a single node, right? And I assume a non-cluster 
setup as H K didn't mention it.

Regards,
O.

> > This was not a problem using Sling 7 with Jackrabbit (instead of Oak) and
> > connecting over RMI.
> That was because there you were not using a cluster. You were
> connecting to the JR running within Sling using RMI. So it was a
> single node setup. If you use RMI now also and connect to same Sling
> node then it would work as expected.
> 
> Chetan Mehrotra
> [1] http://jackrabbit.apache.org/oak/docs/architecture/nodestate.html

Reply via email to