hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Shumin Guo <gsmst...@gmail.com>
Subject Re: how to save dead node when it's disk is full?
Date Thu, 29 May 2014 03:22:40 GMT
If the datanode is dead, block replications will be re-replicated on other
datanodes automatically. So, after a while, when those blocks have been
re-replicated, you should be able to delete the data on the dead nodes and
restart it. Here, I am assuming your cluster have enough space on other
live nodes for the block replications.


On Wed, May 28, 2014 at 7:34 PM, ch huang <justlooks@gmail.com> wrote:

> hi,maillist:
>                  one of my datanode is full ,so dead in cluster ,how can i
> do to let it live again?
>

Mime
View raw message