Thanks, Val. I think it is worthwhile to change the ticket description with
full context at this point. Also, the title of the ticket should not have
CacheStore, but CacheConfiguration. Can you do it?
On Tue, Feb 21, 2017 at 2:38 PM, Valentin Kulichenko <
[hidden email]> wrote:
> Dmitry,
>
> I believe you're talking about this issue:
>
https://issues.apache.org/jira/browse/IGNITE-1903>
> I just added a comment there with some thoughts about how we can improve
> this.
>
> -Val
>
> On Mon, Feb 20, 2017 at 12:37 AM, Alexey Goncharuk <
>
[hidden email]> wrote:
>
> > Dmitriy,
> >
> > What do you mean by deploying cache metadata? The node filter itself must
> > be deserialized on each node because we need to evaluate it. As for the
> > CacheConfiguration, we do not need to deserialize the configuration on
> > every node, this will be a good change for Ignite 2.0.
> >
> > 2017-02-16 23:14 GMT+03:00 Dmitriy Setrakyan <
[hidden email]>:
> >
> > > Igniters,
> > >
> > > I believe that we are not propagating cache metadata correctly, but I
> > want
> > > to confirm it with the community first.
> > >
> > > Currently, it is possible in Ignite to split the cluster into logical
> > > groups, and deploy caches only on certain nodes using NodeFilters. For
> > > example, CacheA will reside only on Nodes 1, 2, and 3, and CacheB will
> > > reside only on nodes 4 and 5. However, even in this case, it looks like
> > the
> > > cache metadata for both caches will be deployed on all the nodes.
> > >
> > > It looks like the correct approach would be to honor the NodeFilter
> for a
> > > cache whenever deploying the cache metadata. Is there a reason for the
> > > metadata to be deployed on all the nodes?
> > >
> > > D.
> > >
> >
>
Locked
