hadoop-hdfs-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Todd Lipcon <t...@cloudera.com>
Subject Re: Source Code Question about blockReport
Date Tue, 22 Jun 2010 17:05:33 GMT
On Mon, Jun 21, 2010 at 10:59 PM, Jeff Zhang <zjffdu@gmail.com> wrote:

> Hi Hadoop Devs,
>
> I have one question about the blockReport DataNode send to NameNode. I
> think NameNode get the blockReport from DataNode, then it can tell
> DataNode which block is invalid and which block should be replicated,
> But I look at the source code of method blockReport of NameNode, it
> always returns DatanodeCommand.FINALIZE or null. it makes me confused
> here. Could anybody help me understand what's the meaning of the code
> here ? And how does the NameNode tell DataNode which block should be
> delete ? Thanks for your help.
>

Those commands come in the DatanodeCommand in response to sendHeartbeat.

-Todd


>
>
> --
> Best Regards
>
> Jeff Zhang
>



-- 
Todd Lipcon
Software Engineer, Cloudera

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message