Hi Dan,

I see - thanks for the detailed explanation!

In the spirit of modularity and cleaner dependencies, the core should not 
depend on any custom integrations on application (Spark, Redis, Lucene) or 
storage side (HDFS, HBase?). I also see hbase jar within 'apache-geode/lib' 
directory. Do you know what that is used for?

To answer your question, such dependencies should ideally be pulled-in only at 
runtime or better be configurable. Seems Zeppelin has a concept of profiles 
that builds 'integration-specific' artifacts.

I'm sure this is easier said than done but that's a starting view...

Thanks,
Nitin

________________________________________
From: Dan Smith <[email protected]>
Sent: Friday, February 19, 2016 2:09 PM
To: geode
Subject: Re: Lucene function added by default

Hi Nitin,

I think this is a bug that the function shows up in that list. We actually
have a lot of other internal functions that don't show up in that list
because they implement the "InternalEntity" marker interface. So this
function should probably implement that as well.

There's another question this brings up, which is whether the lucene
integration and lucene jars should be added by default to the classpath
when you start a gemfire server. That I'm not as sure about. Thoughts?

-Dan

On Fri, Feb 19, 2016 at 1:48 PM, Nitin Lamba <[email protected]> wrote:

> Hi,
>
>
> While starting-up a new geode cluster, I see that a lucene function is
> registered even before any regions are created. Following are the simple
> steps to reproduce it:
>
>
> ./bin/gfsh
>
> gfsh>start locator --name=Locator
>
> gfsh>start server --name=Server1
>
> gfsh>list functions
>
>
> Member  | Function
>
> ------- |
> ---------------------------------------------------------------------
>
> Server1 |
> com.gemstone.gemfire.cache.lucene.internal.distributed.LuceneFunction
>
>
> Is this expected/ necessary? I would expect such a function to be defined
> only when Lucene integration is intended to be used. Please advise if a
> JIRA should be opened for this.
>
>
> Thanks,
>
> Nitin
>
>
>

Reply via email to