You are viewing a plain text version of this content. The canonical link for it is here.

Posted to solr-user@lucene.apache.org by Otis Gospodnetic <ot...@gmail.com> on 2013/05/17 01:34:48 UTC

Controlling which node(s) hold(s) a collection

Hi,

Is it possible to control on which node(s) a collection should be placed?

I've looked at http://wiki.apache.org/solr/SolrCloud and
http://wiki.apache.org/solr/CoreAdmin and have searched the ML
archives, but couldn't find any mentions of that.

Use case:
* Want to use SolrCloud for large indices that I want to shard and replicate
* Have a number of smaller indices that need to live in the same
cluster, but that I don't want to shard - queries are fast when
executed against the whole index being on a single server, and they
use join and pivot faceting, neither of which works with sharded
indices

I have 30+ such non-shardable indices of varying sizes and I want to
make sure they are distributed over all cluster nodes nice and evenly.
 I'm assuming there is no better way than to manually control
placement of my 1-shard collections (iffff that's even doable), but if
there is a better way, I'm all eyeballs!

Thanks,
Otis
--
Search Analytics - http://sematext.com/search-analytics/index.html
Performance Monitoring - http://sematext.com/spm/index.html

Re: Controlling which node(s) hold(s) a collection

Posted by Otis Gospodnetic <ot...@gmail.com>.

Thanks Mark.

> The collections api also takes a list of nodes names to use optionally.

I just tried looking for this, but I don't see it on
http://wiki.apache.org/solr/SolrCloud - either I need new glasses or
it's not mentioned there?

Thanks,
Otis
--
Search Analytics - http://sematext.com/search-analytics/index.html
SOLR Performance Monitoring - http://sematext.com/spm/index.html




On Thu, May 16, 2013 at 7:44 PM, Mark Miller <ma...@gmail.com> wrote:
> You can control simply with the CoreAdmin api - the core is created at the location of whatever url you use…simply fire the creates at whatever nodes you want the collection to live on.
>
> The collections api also takes a list of nodes names to use optionally.
>
> - Mark
>
> On May 16, 2013, at 7:34 PM, Otis Gospodnetic <ot...@gmail.com> wrote:
>
>> Hi,
>>
>> Is it possible to control on which node(s) a collection should be placed?
>>
>> I've looked at http://wiki.apache.org/solr/SolrCloud and
>> http://wiki.apache.org/solr/CoreAdmin and have searched the ML
>> archives, but couldn't find any mentions of that.
>>
>> Use case:
>> * Want to use SolrCloud for large indices that I want to shard and replicate
>> * Have a number of smaller indices that need to live in the same
>> cluster, but that I don't want to shard - queries are fast when
>> executed against the whole index being on a single server, and they
>> use join and pivot faceting, neither of which works with sharded
>> indices
>>
>> I have 30+ such non-shardable indices of varying sizes and I want to
>> make sure they are distributed over all cluster nodes nice and evenly.
>> I'm assuming there is no better way than to manually control
>> placement of my 1-shard collections (iffff that's even doable), but if
>> there is a better way, I'm all eyeballs!
>>
>> Thanks,
>> Otis
>> --
>> Search Analytics - http://sematext.com/search-analytics/index.html
>> Performance Monitoring - http://sematext.com/spm/index.html
>

Re: Controlling which node(s) hold(s) a collection

Posted by Mark Miller <ma...@gmail.com>.

You can control simply with the CoreAdmin api - the core is created at the location of whatever url you use…simply fire the creates at whatever nodes you want the collection to live on.

The collections api also takes a list of nodes names to use optionally.

- Mark

On May 16, 2013, at 7:34 PM, Otis Gospodnetic <ot...@gmail.com> wrote:

> Hi,
> 
> Is it possible to control on which node(s) a collection should be placed?
> 
> I've looked at http://wiki.apache.org/solr/SolrCloud and
> http://wiki.apache.org/solr/CoreAdmin and have searched the ML
> archives, but couldn't find any mentions of that.
> 
> Use case:
> * Want to use SolrCloud for large indices that I want to shard and replicate
> * Have a number of smaller indices that need to live in the same
> cluster, but that I don't want to shard - queries are fast when
> executed against the whole index being on a single server, and they
> use join and pivot faceting, neither of which works with sharded
> indices
> 
> I have 30+ such non-shardable indices of varying sizes and I want to
> make sure they are distributed over all cluster nodes nice and evenly.
> I'm assuming there is no better way than to manually control
> placement of my 1-shard collections (iffff that's even doable), but if
> there is a better way, I'm all eyeballs!
> 
> Thanks,
> Otis
> --
> Search Analytics - http://sematext.com/search-analytics/index.html
> Performance Monitoring - http://sematext.com/spm/index.html