lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erick Erickson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SOLR-4011) SolrCore - correct behavior while having error loading specific cores
Date Mon, 29 Oct 2012 12:16:15 GMT

    [ https://issues.apache.org/jira/browse/SOLR-4011?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13485983#comment-13485983
] 

Erick Erickson commented on SOLR-4011:
--------------------------------------

I'm not at all sure we _should_ fix this. In the vast majority of cases, just continuing would
lead to hard-to-understand behavior. Imagine an innocent typo on a core definition line in
the non-puppet (or other "special") setup for just one of your cores. Now you have users who
can't find a document, but only when searching for that core, but you don't know why. The
"fail fast" option makes that obvious when you start solr, not after you've deployed it to
your production system and have customers complaining....

You can actually fire up a Solr with _no_ cores defined, and then use the "CREATE" command
after you're sure the relevant file structure is in place. If you combine that with "persist=true"
in your solr.xml file, you're in a situation where after you've deployed all your infrastructure
and issued the proper "CREATE" commands, you'll have the setup in persisted to your solr.xml
file. Despite its name, CREATE can be used as though it were something like OPEN or LOAD on
a pre-existing collection.

In the future, could you raise this kind of question on the user's list first rather than
opening a JIRA? We try to keep JIRAs for confirmed bugs...

Best
Erick


                
> SolrCore - correct behavior while having error loading specific cores
> ---------------------------------------------------------------------
>
>                 Key: SOLR-4011
>                 URL: https://issues.apache.org/jira/browse/SOLR-4011
>             Project: Solr
>          Issue Type: Improvement
>          Components: multicore
>         Environment: RHEL 6.1/6.3
> Solr 3.6.1
>            Reporter: Gianluca Varisco
>            Priority: Minor
>
> I've a multicore setup.  Solr.xml contains the following:
> <solr persistent="false">
>   <!--
>   adminPath: RequestHandler path to manage cores.  
>     If 'null' (or absent), cores will not be manageable via request handler
>   -->
> <cores adminPath="/admin/cores">
> <core name="a" instanceDir="/shop/www/htdocs/a/shop.staging/solr/app" dataDir="/shop/solr/staging/a/data/"
configName="solrconfig.xml" schemaName="schema.xml" />
> <core name="b" instanceDir="/shop/www/htdocs/b/shop.staging/solr/app" dataDir="/shop/solr/staging/b/data/"
configName="solrconfig.xml" schemaName="schema.xml" />
> <core name="c" instanceDir="/shop/www/htdocs/c/shop.staging/solr/app" dataDir="/shop/solr/staging/c/data/"
configName="solrconfig.xml" schemaName="schema.xml" />
> </cores>
> </solr>
> When solr starts, it tries to LOAD all the cores. If one fails (no solrconfig.xml or
instanceDir found), the whole multicore configuration is broken and Solr won't answer as well
for the other two cores. 
> Is there a workaround for it? Reasons while it can happen is, for example, if cores'
creation in the XML are managed indipendently (e.g via Puppet) but applications are not deployed
in that instanceDir yet (therefore: no solrconfig.xml/schema.xml). Is there a specific parameter
to use?
>  

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

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


Mime
View raw message