lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mark Miller (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SOLR-8856) Do not cache merge or 'read once' contexts in the hdfs block cache.
Date Mon, 21 Mar 2016 17:53:25 GMT

    [ https://issues.apache.org/jira/browse/SOLR-8856?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15204739#comment-15204739
] 

Mark Miller commented on SOLR-8856:
-----------------------------------

bq. What is our compatibility promise? Do we need to maintain the old constructor that is
no longer being used?

Solr does not promise back compat on low level Java API's. Just HTTP and our best effort on
solrj APIs.

But also, this is a patch for trunk which has no back compat promises.

bq. maybe it is time to use a builder or a configuration object? 

Not really worth it IMO - this is not supposed to be a user facing class in Solr and I don't
see it being instantiated somewhere else anytime soon either.

> Do not cache merge or 'read once' contexts in the hdfs block cache.
> -------------------------------------------------------------------
>
>                 Key: SOLR-8856
>                 URL: https://issues.apache.org/jira/browse/SOLR-8856
>             Project: Solr
>          Issue Type: Improvement
>            Reporter: Mark Miller
>            Assignee: Mark Miller
>         Attachments: SOLR-8856.patch, SOLR-8856.patch, SOLR-8856.patch
>
>
> Generally the block cache will not be large enough to contain the whole index and merges
can thrash the cache for queries. Even if we still look in the cache, we should not populate
it.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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


Mime
View raw message