hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "DENG FEI (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (HDFS-11664) Invalidate blocks cannot invoke because of replication work slowly while cluster is almost full
Date Mon, 24 Apr 2017 03:19:04 GMT

    [ https://issues.apache.org/jira/browse/HDFS-11664?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15980679#comment-15980679
] 

DENG FEI edited comment on HDFS-11664 at 4/24/17 3:18 AM:
----------------------------------------------------------

Normally should maintain definite  storage remaining,but if out of control,many blocks need
to do replication cause of writing minimum replica,althought immediatelly delete heavy path
and file,the cluster can't recovery for a long time.


was (Author: deng fei):
Normally should maintain definite  storage remaining,but if out of control,many blocks need
to do replication cause of write minimum replica,althought immediatelly delete heavy path
and file,the cluster can't recovery for a long time.

> Invalidate blocks cannot invoke because of replication work slowly while cluster is almost
full
> -----------------------------------------------------------------------------------------------
>
>                 Key: HDFS-11664
>                 URL: https://issues.apache.org/jira/browse/HDFS-11664
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>          Components: namenode
>    Affects Versions: 2.7.2, 3.0.0-alpha2
>            Reporter: DENG FEI
>         Attachments: HDFS-11664.001.patch
>
>
> Currently invalideBlock is executed after replication, when 'dfs remaining' is not much
left and affect block writing, but can not release sufficient storage quickly after did delete,
the daemon stuck at computing replication.
> Split replication and invalidation may be better.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
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