You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "Yi Liu (JIRA)" <ji...@apache.org> on 2015/03/10 15:12:39 UTC
[jira] [Commented] (HDFS-7846) Create off-heap BlocksMap and
BlockData structures
[ https://issues.apache.org/jira/browse/HDFS-7846?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14354936#comment-14354936 ]
Yi Liu commented on HDFS-7846:
------------------------------
Thanks Colin for the patch, it looks good. Each shard includes a blocks htable, it's great we use separate lock for the shard. Then we don't need to use FSN lock.
I have one question and few comments about code.
For original BlockInfo, it includes triplets which let us easily get blocks belonging to some DatanodeStorage. In this patch, seems not. We will use a different approach in other JIRAs? Maybe I need to wait for HDFS-7848.
Few nits:
*1.* In {{BlockMap#Builder#build}}, duplicate check for {{this.mman != null}}
*2.* Please add InterfaceAudience and InterfaceStability to public class
*3.* unneccessary import in new added class files
> Create off-heap BlocksMap and BlockData structures
> --------------------------------------------------
>
> Key: HDFS-7846
> URL: https://issues.apache.org/jira/browse/HDFS-7846
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Affects Versions: HDFS-7836
> Reporter: Colin Patrick McCabe
> Assignee: Colin Patrick McCabe
> Attachments: HDFS-7846-scl.001.patch
>
>
> Create off-heap BlocksMap, BlockInfo, and DataNodeInfo structures. The BlocksMap will use the off-heap hash table.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)