hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Allen Wittenauer (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-10999) Use more generic "low redundancy" blocks instead of "under replicated" blocks
Date Thu, 13 Oct 2016 16:56:20 GMT

    [ https://issues.apache.org/jira/browse/HDFS-10999?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15572492#comment-15572492
] 

Allen Wittenauer commented on HDFS-10999:
-----------------------------------------

Given that the fsck output is the *only* way to get some pieces of information, changing fsck
is almost always a major, ops breaking event.  Coupling that with pretty much breaking metrics
collection... There is no "supposed" here: this is very very incompatible and will cause admins
to burn Apache Hadoop conferences to the ground in their anger if we aren't careful.

That said, I can empathize with the EC folks.  'under replicated' doesn't really cover the
state of a block with a missing reconstructable chunk.  But I'm not sure that 'low redundancy'
necessarily conveys the state of a non-EC block either. If I'm not running EC at all, it comes
across as a gratuitous change.  I need to think more about this, to be honest.

But some questions first:

I'm trying to remember, is it possible to have EC and non-EC blocks in a file system?  If
not, what about in the future? Are we actually trying to shoe horn two separate measurements
into the same metric here?  Is there a situation where having both under replicated and low
redundancy blocks makes sense?  How does the storage policy interact with a change like this?

> Use more generic "low redundancy" blocks instead of "under replicated" blocks
> -----------------------------------------------------------------------------
>
>                 Key: HDFS-10999
>                 URL: https://issues.apache.org/jira/browse/HDFS-10999
>             Project: Hadoop HDFS
>          Issue Type: Sub-task
>          Components: erasure-coding
>    Affects Versions: 3.0.0-alpha1
>            Reporter: Wei-Chiu Chuang
>            Assignee: Yuanbo Liu
>              Labels: supportability
>
> Per HDFS-9857, it seems in the Hadoop 3 world, people prefer the more generic term "low
redundancy" to the old-fashioned "under replicated". But this term is still being used in
messages in several places, such as web ui, dfsadmin and fsck. We should probably change them
to avoid confusion.
> File this jira to discuss it.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org


Mime
View raw message