hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Purtell (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-10655) ByteBuf-fronted slab allocator for on- and off-heap cellblock storage
Date Mon, 03 Mar 2014 03:41:22 GMT

    [ https://issues.apache.org/jira/browse/HBASE-10655?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13917714#comment-13917714

Andrew Purtell commented on HBASE-10655:

On HBASE-10191, [~mcorgan] suggested the allocator could also do constant small (sequential)

We can maintain new metrics on numbers of buffers allocated, stats on arenas, stats on wastage
and internal fragmentation of the buffers, etc, and use these to guide additional optimizations
and refinements.

> ByteBuf-fronted slab allocator for on- and off-heap cellblock storage
> ---------------------------------------------------------------------
>                 Key: HBASE-10655
>                 URL: https://issues.apache.org/jira/browse/HBASE-10655
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Andrew Purtell
>            Assignee: Andrew Purtell
> Our ByteRange interface has some nice properties such as a compareTo method using hbase-friendly
unsigned byte comparison. This can be backed by either on-heap or off-heap byte buffers. Consider
an allocator that internally allocates huge on or off-heap slabs and hands out ByteRanges
over portions of those arenas. Keep it as simple as possible, optimized for HBase needs. Can
be fairly standalone with respect to the rest of the HBase code. 

This message was sent by Atlassian JIRA

View raw message