[infinispan-dev] How the remote fetching for GET works in DIST mode

[Dan Berindei]

Yep, this was discussed before, see
http://markmail.org/thread/isqzt77ywclokedd (it starts with other
stuff, but it discusses this as well).

There's also an issue in JIRA: https://issues.jboss.org/browse/ISPN-825

Cheers
Dan


On Thu, Apr 26, 2012 at 1:52 AM, Michal Linhard <mlinhard at redhat.com> wrote:
> Maybe this is an ancient topic discussed, ages ago somewhere, but if
> there's a quick answer, please save me from searching through tons of
> archives.
>
> Let's have a DIST mode cache.
>
> When we're doing a memcached/REST tests where for each request there's
> cca numOwners/numNodes probability that we'll hit the owner,
> in numNodes=4, numOwners=2 this means cca every second request needs a
> remote fetch (assume no L1 cache)
>
> I'm just looking at the DistributionManagerImpl.retrieveFromRemoteSource
> and trying understand how it works.
>
> Does it really cause 2 more GET request processing in the cluster ?
> OK even when we're assuming serving a GET request should be quick, I'd
> like to see how it would perform if we only allowed getting from the
> first owner.
> I guess right now we don't allow configuring this. Are there any catches
> with this ?
>
> m.
>
> --
> Michal Linhard
> Quality Assurance Engineer
> JBoss Datagrid
>
> Red Hat Czech s.r.o.
> Purkynova 99 612 45 Brno, Czech Republic
> phone: +420 532 294 320 ext. 62320
> mobile: +420 728 626 363
>
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev