lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Erick Erickson <>
Subject Re: (ANNOUNCEMENT) Solr Examples reading group
Date Thu, 06 Oct 2016 02:06:32 GMT
I'm a little reluctant to remove id and _version_ as they're
so essential to SolrCloud. Stand alone for that matter, I
strongly recommend that people have a <uniqueKey>
no matter what as the "usual" expectation is to replace docs
I think...

But perhaps a minimal_config and milnimal_cloud_config?

In either case I think <unkqueKey> should be there.....


On Wed, Oct 5, 2016 at 9:58 AM, Alexandre Rafalovitch
<> wrote:
> does Solr as part of Big Data, but not sure about Nutch
> has Solr and Nutch, but not
> sure about UI
> LucidWorks' commercial Fusion platform I believe has both with UI
> Solr default config has LOTS of comments :-) Of course, the ones in
> managed-schema go away on first "schemaless mode" modification, so
> read them before they go go go. And for solrconfig.xml, if you
> reformat - say "films" - example to remove comments and spaces, you
> drop from 1482 lines to 271.
> So, I should say comments are not a problem. At least in the "not
> enough" sense. In the "drown in" sense, that's a different story.
> Regards,
>    Alex.
> ----
> Newsletter and resources for Solr beginners and intermediates:
> On 5 October 2016 at 23:26, WebDawg <> wrote:
>> On Wed, Oct 5, 2016 at 11:15 AM, Alexandre Rafalovitch
>> <> wrote:
>>> I have an nearly minimal config for Solr 4.9, 5.5 and 6.2 at:
>>> . They are _quite_
>>> feature minimal (6.2 does not even have ID or version), but they will
>>> import and search example/exampledocs/*.xml without fuss
>>> I challenge you to guess number of lines they take before looking :-)
>>> As to minimal examples, I think that's absolutely the next step. Did
>>> you ever notice that we don't even get too many question on the Solr
>>> Users mailing list about the example - I suspect it is because people
>>> get lost so fast they don't even know how to ask the questions.
>>> SOLR-9601 will be my next experiment for that.
>>> As to kitchen-sync vs the examples, maybe we need a "configuration"
>>> configset that will have all the definitions and then have
>>> Admin/Velocity/whatever UI that allows user pick a definition and
>>> existing core on the same server and it will the
>>> type/field/configuration command to that core. It is definitely doable
>>> for schema. For solrconfig it is a bit more challenging as not
>>> everything is API controlled yet and API itself - in my mind - is
>>> still very raw. The support files is a small challenge too for
>>> non-cloud mode. But this could be the dogfood project to fix that.
>>> So, as a workflow, it would be:
>>> 1) bin/solr create -c mycore -d minimal_config
>>> 2) Experiment to the limit of that core's configuration (initial types, etc)
>>> 3) bin/solr create -c config -d configuration
>>> 4) Use the new core's enhanced UI to do whatever is needed to mycore
>>> 5) bin/solr delete -c config
>>> 6) Continue working with mycore
>>> The other similar idea is a 3rd party solr config wizard that walks
>>> the user through the choices and generates config files with/without
>>> comments/links to ref guide. Then, it is not limited to Solr file
>>> layout. On the other hand, people need to know the tool exists....
>>> Regards,
>>>    Alex.
>>> P.s. I will try very hard to be at the meetup on Tuesday, so will be
>>> very happy to have an in-depth discussion about this topic
>> I was thinking about working on something like this but I just asked
>> the nutch list if anything like this existed.
>> Does it?  Is there anyway to configure lets say SOLR and nutch with a GUI util?
>> I remember back in the day configuring eggdrops and I was very happy
>> that everything was there to comment out if needed.  It made
>> understanding configuration easy.
>> What do the SOLR default configs look like right now?  Are there comments?

View raw message