cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Ellis (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CASSANDRA-408) Pool BufferedRandomAccessFile objects used by sstable reads
Date Wed, 02 Sep 2009 16:04:34 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-408?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12750499#action_12750499
] 

Jonathan Ellis commented on CASSANDRA-408:
------------------------------------------

this results in not one pooled reader per thread but one pooled reader per sstable per thread.
 this would be bad in pathological cases like digg's 1200 sstables post-bulk-load-pre-compaction.
 need to rethink the approach here.

> Pool BufferedRandomAccessFile objects used by sstable reads
> -----------------------------------------------------------
>
>                 Key: CASSANDRA-408
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-408
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: Core
>            Reporter: Jonathan Ellis
>            Assignee: Jonathan Ellis
>             Fix For: 0.5
>
>         Attachments: 408.patch, commons-pool-1.5.2.jar
>
>
> not only does BRAF per op do a whole lot of extra fopens, but the buffering actually
makes it _more_ expensive to set up since on the jvm all primitive arrays are initialized
to zero.
> this adds a simple read test to stress.py; I'm seeing about a 10% increase in throughput
which is worth 200loc imo.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message