[
https://issues.apache.org/jira/browse/CASSANDRA-1945?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12980204#action_12980204
]
Vijay commented on CASSANDRA-1945:
----------------------------------
1) yes if we use java reflection API but don't we have the rows stored in bytes
already?
2) right now there is no way to have a big cache for a certian key space...
even though we provide MMap there is no way to predict the IO's as you
mentioned earlier... and it is not tunable per key space (is it?)... And the
cache with less GC will always help in these cases (Atleast till we get GC
First in jdk 7).... And we also provide choices to the user :)
> Use EHcache for row cache + disk spillover
> ------------------------------------------
>
> Key: CASSANDRA-1945
> URL: https://issues.apache.org/jira/browse/CASSANDRA-1945
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: T Jake Luciani
> Assignee: Vijay
> Attachments: 0001-Config-1945.txt, 0002-Update_existing-1945.txt,
> 0003-New_Cache_Providers-1945.txt, BB_Cache.png, ehcache-1.6.2.jar,
> EhCache-OSS.png, JMX-Cache.png, JMX-EhCache.png, Old_Cahce.png
>
>
> Ehcache does a nice job of spilling over to disk and its ASL2
> http://www.ehcache.org/license.html
> Also, the commercial version supports Smaller JVM sizes via it's BigMemory
> product
> http://www.terracotta.org/bigmemory
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
