So I have been playing around with no locator and the clients directly connecting to the server (stand alone process). So far so good, I have not put it through its paces yet though.
Honestly I am pretty excited about this approach. What about making this an official project under Geode? Basically a pre configured boot app that can a config can be passed to on start up. On Mon, Aug 17, 2015 at 10:48 AM, Lyndon Adams <[email protected]> wrote: > I have seen this model used a few years back. Either use a client with no > locator or fire up a embedded cache without locator. > > Sent from my iPhone > > On 17 Aug 2015, at 16:37, james bedenbaugh <[email protected]> wrote: > > Luke, > > Interesting. I am so used to thinking of Geode as an Enterprise framework. > Are you thinking of this concept in terms of a external small cache and not > embedded like a peer-to-peer without the peer?? - And is I think a locator > is not useful for a single node as pointed out earlier. > > Thinking some more, why not abandon a Geode Java client and use REST > instead? > > On Sat, Aug 15, 2015 at 12:25 PM, Luke Shannon <[email protected]> > wrote: > >> Hi All, >> >> Curious about what everyone thinks about usage of Geode as a single >> process rather than a full cluster. Before you respond to that alone, lets >> me review why I would want to do this :-) >> >> I like the Geode programming model for CRUD operations, function >> executions and listeners. Its CacheWriter and Reader are also really >> useful. I find the Client/Server approach really powerful, interests, >> client side listeners and expiration provide some really powerful features >> for powerful client applications. >> >> So lets say I want all of this but don't need a distributed system (for a >> smaller website lets say). I also don't want to mess with GFSH and making >> any changes at the OS level. I just want something I can start. >> >> I obvious thought was to use Redis, but I wanted to see if I could do >> something with Geode as I am already pretty familiar with it. >> >> As an experiment I built Spring Boot application with an embedded Locator >> and Server (sample config below) that contains the Server config and any >> dependancies my functions and listeners needed. Whats nice here is I have a >> jar file I can copy somewhere, start up and be instantly ready for a client >> to connect too. I have 4 clients and they get fast responses to Key/Value >> operations, execute functions, receive interests, etc. I monitor it with >> Monit. >> >> Although I have not tried, I am pretty sure I can even run it on >> run.pivotal.io. >> >> Thoughts on this approach? Should I really just be using Redis for a >> single cache? >> >> Snippet from cache-config.xml >> >> <util:properties id="singleCacheConfigurationSettings"> >> >> <prop key="name">singleCache</prop> >> >> <prop key="locators">127.0.0.1[11235]</prop> >> >> <prop key="log-level">config</prop> >> >> <prop key="mcast-port">0</prop> >> >> <prop key="start-locator">127.0.0.1[11235]</prop> >> >> </util:properties> >> >> <gfe:cache id="gemfireCache" pdx-serializer-ref= >> "reflection-pdx-serializer" >> >> properties-ref="singleCacheConfigurationSettings" /> >> >> <gfe:cache-server port="0" cache-ref="gemfireCache" /> >> > > > > -- > Regards, > Jim Bedenbaugh > Advisory Data Engineer > Pivotal Software > > -- Luke Shannon | Sr. Field Engineer - Toronto | Pivotal ------------------------------------------------------------------------- Mobile:416-571-9495 Join the Toronto Pivotal Usergroup: http://www.meetup.com/Toronto-Pivotal-User-Group/
