You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by Timothy Potter <th...@gmail.com> on 2014/08/28 20:49:59 UTC

Issue with bin/solr script, collection1, and cloud mode

Hi,

Wanted to bring up an issue
(https://issues.apache.org/jira/browse/SOLR-6447) just to get opinions
on whether this warrants a re-spin (thinking NO but wanted to be sure
someone didn't feel otherwise).

Basically, if you use: bin/solr -c, Solr will start in cloud mode and
since collection1 defines a cores.properties, it gets auto-created as
a collection. Unfortunately, the script doesn't set -DnumShards, so
collection1 is getting created using the implicit router and a null
range.

As for overall impact, I think it's pretty minor but could lead to
confusion for new users that may want to try splitting collection1
(unlikely).

For now, I've updated the Ref Guide to guide new users to using:
bin/solr -e cloud instead, which will prompt the user for number of
nodes, collection name, num shards, rf, etc. In other words, if users
do: bin/solr -e cloud, it will help them create a new collection that
has numShards set correctly.

I don't think there's any big risk here but wanted to be proactive in
letting folks know this issue exists before we move forward with 4.10.

Cheers,
Tim

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Re: Issue with bin/solr script, collection1, and cloud mode

Posted by "david.w.smiley@gmail.com" <da...@gmail.com>.
Ok.

I wish the router was an explicit option, separate from declaring
numShards.  And furthermore, that it would never be “implicit” unless you
expressly told it to be.  People sometimes get this router because they
forget numShards, thinking “1 is fine anyway”. — for now.

~ David Smiley
Freelance Apache Lucene/Solr Search Consultant/Developer
http://www.linkedin.com/in/davidwsmiley


On Thu, Aug 28, 2014 at 2:49 PM, Timothy Potter <th...@gmail.com>
wrote:

> Hi,
>
> Wanted to bring up an issue
> (https://issues.apache.org/jira/browse/SOLR-6447) just to get opinions
> on whether this warrants a re-spin (thinking NO but wanted to be sure
> someone didn't feel otherwise).
>
> Basically, if you use: bin/solr -c, Solr will start in cloud mode and
> since collection1 defines a cores.properties, it gets auto-created as
> a collection. Unfortunately, the script doesn't set -DnumShards, so
> collection1 is getting created using the implicit router and a null
> range.
>
> As for overall impact, I think it's pretty minor but could lead to
> confusion for new users that may want to try splitting collection1
> (unlikely).
>
> For now, I've updated the Ref Guide to guide new users to using:
> bin/solr -e cloud instead, which will prompt the user for number of
> nodes, collection name, num shards, rf, etc. In other words, if users
> do: bin/solr -e cloud, it will help them create a new collection that
> has numShards set correctly.
>
> I don't think there's any big risk here but wanted to be proactive in
> letting folks know this issue exists before we move forward with 4.10.
>
> Cheers,
> Tim
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: dev-help@lucene.apache.org
>
>