hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jing Zhao (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-7716) Erasure Coding: extend BlockInfo to handle EC info
Date Tue, 10 Feb 2015 19:08:12 GMT

    [ https://issues.apache.org/jira/browse/HDFS-7716?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14314685#comment-14314685

Jing Zhao commented on HDFS-7716:

Thanks for the comments, Zhe!

bq. What if we are removing an over-replicated slot? I guess we should still swap it with
the last node, to avoid growing triplets too much?

This will not cause triplets growing too much because the empty slots after {{dataBlockNum
+ parityBlockNum}} will be reused when adding new over-replicated storages. We can leave the
empty slots there or we can choose to shrink the triplets array. Currently I just keep the
empty slots to avoid copying the whole array.

bq. This is to guarantee indices has the same size as triplets?

Actually this is to make sure the {{newBlock}} has enough size to store all the triplets (including
over-replicated blocks).

> Erasure Coding: extend BlockInfo to handle EC info
> --------------------------------------------------
>                 Key: HDFS-7716
>                 URL: https://issues.apache.org/jira/browse/HDFS-7716
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>            Reporter: Jing Zhao
>            Assignee: Jing Zhao
>         Attachments: HDFS-7716.000.patch, HDFS-7716.001.patch
> The current BlockInfo's implementation only supports the replication mechanism. To use
the same blocksMap handling block group and its data/parity blocks, we need to define a new
BlockGroupInfo class.

This message was sent by Atlassian JIRA

View raw message